Active sources: 7. Sources represented in this queue: 3. The cron runs daily at 06:00 server time; arxiv RSS is often empty on weekends.
Linked to your results
- score 100arxiv cs.AI (Artificial Intelligence)arxiv:2606.19808unread
Think Again or Think Longer? Selective Verification for Budget-Aware Reasoning
Sajib Acharjee Dip, Dawei Zhou, Liqing Zhang · 2026-06-20
arXiv:2606. 19808v1 Announce Type: new Abstract: Test-time reasoning is increasingly used as a serving-time control knob, but extra reasoning is not uniformly valuable: it can repair failed attempts, waste compute on already-correct answers, or introduce harmful answer changes.
Read next because Think Again or Think Longer? Selective Verification for Budget-Aware Reasoning overlaps with clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)", clean result "Coupling evil personas with wrong answers fails to protect Qwen2.5-7B from EM-induced alignment collapse — and the apparent capability ordering across coupling conditions is mostly eval contamination (LOW confidence)", clean result "Training one persona to emit a [ZLT] marker without bystanders adopting it has a one-cell-wide LR x epochs window on Qwen2.5-7B-Instruct (LOW confidence)". Matching terms: rect, correct, token, compare, control, test, model. Source: arxiv cs.AI (Artificial Intelligence).
arXiv:2606.19808v1 Announce Type: new Abstract: Test-time reasoning is increasingly used as a serving-time control knob, but extra reasoning is not uniformly valuable: it can repair failed attempts, waste compute on already-correct answers, or introduce harmful answer changes. We study this as a deployment allocation problem rather than a new-verifier problem. We introduce \sevra, Selective Verification for Reasoning Allocation, a serving-layer controller that decides whether to preserve a frozen solver's initial answer or invoke active verification. Using a frozen Qwen3-4B solver, we log intervention outcomes and train recoverability-aware gates from serving-visible attempt state. On \mathfive, selective verification reaches 76.3\% accuracy, compared with 75.5\% for always verifying, while reducing post-generation tokens by 26.8\% and harmful flips from 2.2\% to 1.0\%. However, an 8,192-token initial solve reaches 76.0\% accuracy with 28\% fewer total model tokens, showing that selective recovery is useful but not the best tested cost frontier. In frozen transfer to \gsm, the selective policy verifies only 3.0\% of examples, improves accuracy from 93.4\% to 94.5\%, and reduces verification tokens by 91.2\% relative to always verifying; again, a longer initial solve matches its accuracy with fewer realized tokens. On CommonsenseQA, always-on verification hurts, while Self-Consistency@5 improves accuracy at about five times the realized token cost. The resulting deployment rule is: tune the initial budget first, then use selective recovery when explicit checks, bounded retries, auditability, or regression-risk control matter.
- score 100arxiv cs.AI (Artificial Intelligence)arxiv:2606.19747unread
A Comparative Study of Pretrained Transformer Models for Quranic ASR: Speech Representations, Label Formats, and Dataset Composition
Nabil Mosharraf Hossain (Greentech Apps Foundation, United Kingdom), Riasat Islam (Greentech Apps Foundation, United Kingdom, Queen Mary University of London, United Kingdom), Unaizah Obaidellah (University of Malaya, Malaysia) · 2026-06-20
arXiv:2606. 19747v1 Announce Type: new Abstract: Quran Automatic Speech Recognition (ASR) aims to convert Quranic recitation into text, enabling applications such as aided memorisation tools and Quranic search engines.
Read next because A Comparative Study of Pretrained Transformer Models for Quranic ASR: Speech Representations, Label Formats, and Dataset Composition overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: strong, text, word, line, rate, extraction, without, full. Source: arxiv cs.AI (Artificial Intelligence).
arXiv:2606.19747v1 Announce Type: new Abstract: Quran Automatic Speech Recognition (ASR) aims to convert Quranic recitation into text, enabling applications such as aided memorisation tools and Quranic search engines. However, existing ASR models often exhibit high Word Error Rates (WER) on user-recited verses and lack full coverage of the Quranic corpus. This paper presents a systematic empirical study of domain-specific fine-tuning of pretrained Transformer-based models for Quranic ASR, using advanced speech feature extraction methods: Wav2Vec2.0, HuBERT, and XLS-R. These models apply self-supervised learning by masking portions of input audio and using Transformer architectures to learn context-aware speech features. The pretrained models are fine-tuned on a filtered Quranic dataset exceeding 870 hours of professional and user recitations. Through comprehensive ablation studies across feature extractors, output label formats, training strategies, and clip durations, we identify the key factors that affect transcription accuracy in this domain. Our best-performing configuration achieves a WER of 0.08 on the EveryAyah subset and 0.11 on the combined EveryAyah+Tarteel setting, representing roughly a five-percentage-point gain over the Citrinet baseline (WER = 0.163) while reducing combined-model training time from 140 hours to 40 hours. Arabic text without diacritics yields the best fine-tuning results, and Wav2Vec2-XLSR-53 provides the strongest overall representation. Future work includes improving dataset quality and developing phoneme-aware models to extract deeper speech feature representations for Tajweed-sensitive applications.
- score 100arxiv cs.AI (Artificial Intelligence)arxiv:2606.19741unread
Interpreting Neural Combinatorial Optimization via Evolving Programmatic Bottlenecks
Haocheng Duan, Yuxin Guo, Jieyi Bi, Anqi Xie, Sirui Li, Yining Ma, Cathy Wu · 2026-06-20
arXiv:2606. 19741v1 Announce Type: new Abstract: Neural Combinatorial Optimization (NCO) achieves strong performance, yet its black-box nature remains a key roadblock to deployment and scientific diagnosis.
Read next because Interpreting Neural Combinatorial Optimization via Evolving Programmatic Bottlenecks overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: strong, text, class, rate, stage, position, model. Source: arxiv cs.AI (Artificial Intelligence).
arXiv:2606.19741v1 Announce Type: new Abstract: Neural Combinatorial Optimization (NCO) achieves strong performance, yet its black-box nature remains a key roadblock to deployment and scientific diagnosis. Standard interpretability tools, such as Concept Bottleneck Models (CBMs), are ill-equipped for NCO, whose decisions are dynamic, state-dependent, and lack proper concept vocabulary definition. To close this gap, we introduce Evolving Programmatic Bottlenecks (EPB), to our knowledge, the first framework for interpreting NCO policies by distilling black-box NCO models into human-readable program portfolios. EPB employs an LLM to autonomously evolve a bank of programs, where each program's per-step action distribution serves as the bottleneck. EPB works through an iterative framework: Block I fixes program bank capacity and introduces a hybrid textual-numerical gradient descent scheme that couples numerical gradients for student router updates and textual gradients for LLM-based program revision; Block II dynamically adapts bank capacity via fault-targeted expansion and redundancy pruning. Extensive experiments demonstrate EPB's effectiveness and broad applicability, where the distilled program portfolios largely match original performance. EPB also reveals that NCO behavior shifts across optimization stages and can be approximated as a composition of classic heuristic variants. Our work advances interpretable NCO and establishes EPB as a promising tool for interpreting sequential decision-making models.
- score 100arxiv cs.AI (Artificial Intelligence)arxiv:2606.19602unread
Configurable Clinical Information Extraction with Agentic RAG: What Works, What Breaks, and Why
Osman Alperen \c{C}inar-Kora\c{s}, Marie Bauer, Sameh Khattab, Merlin Engelke, Moon Kim, Stephan Settelmeier, Shigeyasu Sugawara, Fabian Freisleben, Felix Nensa, Jens Kleesiek · 2026-06-20
arXiv:2606. 19602v1 Announce Type: new Abstract: Patient contexts span hundreds of heterogeneous documents and thousands of structured data points, yet the document-level metadata that AI systems need for retrieval and triage is absent or incomplete.
Read next because Configurable Clinical Information Extraction with Agentic RAG: What Works, What Breaks, and Why overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: text, eval, source, line, extraction, contexts, absent. Source: arxiv cs.AI (Artificial Intelligence).
arXiv:2606.19602v1 Announce Type: new Abstract: Patient contexts span hundreds of heterogeneous documents and thousands of structured data points, yet the document-level metadata that AI systems need for retrieval and triage is absent or incomplete. Standard retrieval-augmented generation fails on this data, mishandling temporal reasoning, cross-document dependencies, and missing metadata. We deploy ACIE (Agentic Clinical Information Extraction) at University Medicine Essen: an on-premise agentic RAG pipeline that reasons over complete patient contexts and grounds every answer in source passages for clinician verification. We quantify the metadata gap, trace the architectural decisions it shaped, and evaluate extraction alongside an independent retrospective lymphoma registry study, in which nuclear-medicine physicians verify every extracted value against its cited sources. Across 7,326 judgments, clinicians accepted 96.5\% of extractions, with per-type acceptance ranging from 80\% to 99\%.
- score 100arxiv cs.AI (Artificial Intelligence)arxiv:2606.19522unread
REVEAL++: Differentiable Phenotypic Grouping for Vision-Language Retinal Modeling of Alzheimer's Disease Risk
Ethan Elio Meidinger, Seowung Leem, Zeyun Zhao, Ruogu Fang · 2026-06-20
arXiv:2606. 19522v1 Announce Type: new Abstract: The retina offers a noninvasive window into neurodegenerative disease, capturing subtle structural patterns associated with a risk of future cognitive decline.
Read next because REVEAL++: Differentiable Phenotypic Grouping for Vision-Language Retinal Modeling of Alzheimer's Disease Risk overlaps with clean result "Coupling evil personas with wrong answers fails to protect Qwen2.5-7B from EM-induced alignment collapse — and the apparent capability ordering across coupling conditions is mostly eval contamination (LOW confidence)", clean result "Only continuous soft prefixes hit both EM axes at once on Qwen-2.5-7B-Instruct: discrete prompt searches split between the alignment objective and the distributional objective, and both discretizations of the soft prefix collapse (MODERATE confidence)", clean result "The marker is a representational handle, not a behavioural one — sharing it between a villain persona and the assistant transfers no misalignment (HIGH confidence)". Matching terms: alignment, soft, eval, line, language, model. Source: arxiv cs.AI (Artificial Intelligence).
arXiv:2606.19522v1 Announce Type: new Abstract: The retina offers a noninvasive window into neurodegenerative disease, capturing subtle structural patterns associated with a risk of future cognitive decline. Vision-language alignment frameworks such as REVEAL have shown that pairing retinal fundus images with structured clinical risk narratives improves early prediction of Alzheimer's disease (AD). A key design choice in these approaches is the use of phenotypic grouping, where individuals with similar risk profiles are treated as multi-positive pairs during contrastive learning. However, existing methods operationalize phenotypic similarity as a discrete construct, relying on hard group assignments that impose rigid supervision and decouple group formation from representation learning. We propose a continuous formulation of phenotypic structure within contrastive learning. Rather than assigning samples to fixed clusters, we model inter-subject similarity as a differentiable weighting function derived from intra-modality embedding similarities in both retinal images and risk profiles. These weights define soft multi-positive relationships through a continuous aggregation operator, enabling graded supervision that reflects the spectrum nature of disease risk. We further introduce a soft-target contrastive objective that jointly learns cross-modal alignment and phenotypic structure in an end-to-end manner. Evaluated on UK Biobank retinal imaging data for incident AD prediction, the proposed framework consistently outperforms discrete group-based contrastive learning and standard vision-language baselines. By treating phenotypic similarity as a learnable, continuous signal rather than a fixed grouping rule, our approach provides a principled and robust foundation for population-scale neurodegenerative risk modeling from multi-modal retinal and clinical data.
- score 100arxiv cs.AI (Artificial Intelligence)arxiv:2606.19509unread
LLM Doesn't Know What It Doesn't Know: Detecting Epistemic Blind Spots via Cross-Model Attribution Divergence on Clinical Tabular Data
Akshat Dasula, Prasanna Desikan, Jaideep Srivastava · 2026-06-20
arXiv:2606. 19509v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly applied to structured clinical data, yet whether they can recognize the limits of their own knowledge on such tasks remains unexplored.
Read next because LLM Doesn't Know What It Doesn't Know: Detecting Epistemic Blind Spots via Cross-Model Attribution Divergence on Clinical Tabular Data overlaps with clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)", clean result "Coupling evil personas with wrong answers fails to protect Qwen2.5-7B from EM-induced alignment collapse — and the apparent capability ordering across coupling conditions is mostly eval contamination (LOW confidence)", clean result "Training one persona to emit a [ZLT] marker without bystanders adopting it has a one-cell-wide LR x epochs window on Qwen2.5-7B-Instruct (LOW confidence)". Matching terms: rect, correct, line, rate, without, does, language, model. Source: arxiv cs.AI (Artificial Intelligence).
arXiv:2606.19509v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly applied to structured clinical data, yet whether they can recognize the limits of their own knowledge on such tasks remains unexplored. We study this question through the lens of cross-model attribution divergence with the goal of reducing epistemic uncertainty for structured tasks, comparing Qwen 2.5 7B and XGBoost on a prediction task via attribution divergence analysis. We report four findings. First, LLM verbalized confidence is epistemically vacuous, it outputs a near-constant (0.856-0.937) regardless of whether accuracy is 49% or 75.3%, tracking prompt format rather than prediction quality. Second, the LLM exhibits an inverse difficulty effect: accuracy drops to 64.8% when XGBoost is 99% correct, but matches XGBoost (73.8% vs. 73.1%) when it is moderately uncertain. Third, few-shot examples and SHAP-derived feature evidence are orthogonal, super-additive interventions: they reduce the Attribution Disagreement Score (ADS) from 1.54 to 0.38 and improve accuracy from 49% to 75.3% without training. Fourth, a cross-model calibrator that determined LLM reliability using attribution divergence signals reduces expected calibration error from 0.254 to 0.080, replacing uninformative verbalized confidence with patient-specific reliability estimates, without accessing model internals or requiring repeated inference. We frame these findings as a cold start problem for LLMs on structured data and outline a path toward genuine epistemic self-awareness.
- score 100arxiv cs.AI (Artificial Intelligence)arxiv:2606.19494unread
Hidden Anchors in Multi-Agent LLM Deliberation
Apurba Pokharel, Ram Dantu · 2026-06-20
arXiv:2606. 19494v1 Announce Type: new Abstract: Multi-agent LLM deliberation, where agents exchange and revise answers over several rounds, is increasingly used to improve reasoning and accuracy, yet how and why it works is rarely modelled.
Read next because Hidden Anchors in Multi-Agent LLM Deliberation overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: strong, class, rect, correct, alone, does, full, test. Source: arxiv cs.AI (Artificial Intelligence).
arXiv:2606.19494v1 Announce Type: new Abstract: Multi-agent LLM deliberation, where agents exchange and revise answers over several rounds, is increasingly used to improve reasoning and accuracy, yet how and why it works is rarely modelled. Such deliberation mirrors how humans reach decisions. As social animals we are pulled both by the group, the herd effect that classical opinion-dynamics models such as DeGroot and Friedkin--Johnsen capture, and by our own internal belief, which they do not. We model multi-agent deliberation as a closed-loop dynamical system in which each agent carries a hidden internal belief, its anchor, that continually pulls its opinion regardless of its neighbours. We show this anchor can be recovered from the deliberation alone, and that it explains a behaviour classical consensus rules forbid: an agent's confidence in the correct answer can climb past where any agent started, escaping the space (convexhull) formed by the initial beliefs. Checking whether the recovered anchor also predicts held-out runs (generalizes) gives a simple test for when a model is truly driven bysuch an anchor. Across three open-weight model families this is a spectrum, not all-or-nothing. All anchors' influence are about equally strongly, but they differ in where the anchor sits, and only when it sits far from the initial opinions does deliberation escape the hull and need the full closed-loop model.
- score 100arxiv cs.AI (Artificial Intelligence)arxiv:2606.19464unread
Deontic Policies for Runtime Governance of Agentic AI Systems
Anupam Joshi, Tim Finin, Karuna Pande Joshi, Lalana Kagal · 2026-06-20
arXiv:2606. 19464v1 Announce Type: new Abstract: Autonomous agentic AI systems driven by Large Language Models (LLMs) introduce a new class of security, privacy, and compliance challenges: an agent that can invoke tools, manipulate data, install software, and coordinate with peer agents across organizational boundaries must be constrained not just by authentication and access control, but by the full structure of enterprise governance.
Read next because Deontic Policies for Runtime Governance of Agentic AI Systems overlaps with clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)", clean result "Only continuous soft prefixes hit both EM axes at once on Qwen-2.5-7B-Instruct: discrete prompt searches split between the alignment objective and the distributional objective, and both discretizations of the soft prefix collapse (MODERATE confidence)". Matching terms: class, under, soft, eval, line, control, full, trained. Source: arxiv cs.AI (Artificial Intelligence).
arXiv:2606.19464v1 Announce Type: new Abstract: Autonomous agentic AI systems driven by Large Language Models (LLMs) introduce a new class of security, privacy, and compliance challenges: an agent that can invoke tools, manipulate data, install software, and coordinate with peer agents across organizational boundaries must be constrained not just by authentication and access control, but by the full structure of enterprise governance. This includes specifying what agents are permitted and prohibited from doing, what they areobliged to do after certain actions (e.g., notify the CISO), under what conditions a standing obligation may be waived, and which rules take precedence when policies conflict. This governance problem exceeds what current policy engines provide. Systems such as XACML, Rego, and Cedar address only the permit/prohibit subset of this governance structure. They do not provide obligation lifecycle management, meta-policy conflict resolution, dispensations that waive obligations in specific circumstances, and ontological reasoning over domain class hierarchies commonly found in applications such as healthcare, cybersecurity, or data privacy. We propose AgenticRei, which realizes key governance requirements such as obligations, dispensations, policy conflict resolutions, and reasoning over policies, as well as the basic permit/prohibit constraints. We use a deontic policy language built on the Rei framework, expressed as OWL (Web Ontology Language) and evaluated at runtime by a high-performance logic engine entirely outside the LLM. The same pipeline governs both tool invocations by the agent and agent-to-agent messages. We show through examples that deontic policies capture governance constraints around security and privacy that mostly cannot be expressed in current production engines. Our approach composes naturally with industry-standard frameworks like A2AS.
- score 100arxiv cs.CL (NLP)arxiv:2606.19815unread
Clusters are All You Need: Pre-Training the Tsetlin Machine with Semantic Clusters from Language Models for Interpretability
Jiechao Gao, Rohan Kumar Yadav, Yuangang Li, Yuandong Pan, Jie Wang, Ying Liu, Michael Lepech · 2026-06-20
arXiv:2606. 19815v1 Announce Type: new Abstract: Pre-trained language models such as BERT achieve strong text classification performance but lack transparency, limiting their use in high-stakes settings.
Read next because Clusters are All You Need: Pre-Training the Tsetlin Machine with Semantic Clusters from Language Models for Interpretability overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: strong, text, word, class, without, full, trained, language. Source: arxiv cs.CL (NLP).
arXiv:2606.19815v1 Announce Type: new Abstract: Pre-trained language models such as BERT achieve strong text classification performance but lack transparency, limiting their use in high-stakes settings. The Tsetlin Machine (TM) offers fully interpretable, clause-based reasoning but captures little semantic information, and prior attempts to bridge the two rely on static word embeddings that miss contextual meaning. We propose a semantic pre-training framework that transfers knowledge from a pre-trained language model into a TM without using embeddings. Text samples are grouped into semantically coherent clusters with K-means or Top2Vec, and the resulting cluster-sample pairs pre-train a non-negated TM with enhanced Type I feedback. The TM thereby learns interpretable semantic keywords that are fine-tuned on downstream tasks. Across five datasets, our method substantially outperforms vanilla and embedding-based TMs and reaches performance competitive with BERT while remaining interpretable.
- score 100arxiv cs.CL (NLP)arxiv:2606.19744unread
Beyond Uniform Forgetting: A Study of Sequential Direct Preference Optimization Across Preference Settings
Pranav Bhandari, Nicolas Fay, Amitava Datta, Usman Naseem, Mehwish Nasim · 2026-06-20
arXiv:2606. 19744v1 Announce Type: new Abstract: Aligning language models with human preferences often requires optimising multiple behavioural objectives.
Read next because Beyond Uniform Forgetting: A Study of Sequential Direct Preference Optimization Across Preference Settings overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: strong, rect, alignment, distributional, eval, line, does, length. Source: arxiv cs.CL (NLP).
arXiv:2606.19744v1 Announce Type: new Abstract: Aligning language models with human preferences often requires optimising multiple behavioural objectives. A practical approach is to apply these objectives sequentially using preference optimisation methods such as Direct Preference Optimisation (DPO), but it remains unclear whether later training uniformly degrades preferences learned earlier or whether the effect depends on the relationship between objectives. We study sequential DPO across four preference settings covering distributional conflict, multi-attribute interaction, strong safety signal, and compatible response-quality objectives. Using Llama-3.1-8B-Instruct with LoRA adapters, we evaluate all objectives after every stage with a fixed base-model reference. We find that sequential DPO does not produce a single forgetting pattern; preference change ranges from partial degradation to stability, pair-level redistribution, or positive transfer depending on objective relationship, signal strength, and training order. Pair-level analysis using length-normalised policy margins shows that aggregate metrics can mask heterogeneous changes across preference pairs, whereas quartile decomposition reveals that high-confidence pairs can either degrade or improve depending on the setting. Mechanistic diagnostics show that Stage~2 gradients and adapter updates are near-orthogonal to the previous objective across all settings, providing little evidence that direct gradient opposition is the primary driver. These findings suggest that future sequential alignment pipelines should account for objective compatibility and signal strength, rather than assuming that later objectives affect earlier preferences uniformly.
- score 100arxiv cs.CL (NLP)arxiv:2606.19710unread
FineREX: Fine-Tuned NER-RE for Human Smuggling Knowledge Graphs
Elijah Feldman, Dipak Meher, Carlotta Domeniconi · 2026-06-20
arXiv:2606. 19710v1 Announce Type: new Abstract: Court proceedings contain valuable evidence about human smuggling networks, but this information is often buried within unstructured, jargon-heavy legal documents.
Read next because FineREX: Fine-Tuned NER-RE for Human Smuggling Knowledge Graphs overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: text, line, rate, extraction, compare, stage, language, model. Source: arxiv cs.CL (NLP).
arXiv:2606.19710v1 Announce Type: new Abstract: Court proceedings contain valuable evidence about human smuggling networks, but this information is often buried within unstructured, jargon-heavy legal documents. While large language models (LLMs) can support knowledge graph construction through automated information extraction, existing approaches rely on general-purpose models that are not tailored to the entity and relationship definitions required in this domain. We introduce FineREX, a streamlined knowledge graph construction pipeline built around a fine-tuned LLM for named entity recognition and relationship extraction (NER-RE). Using a manually annotated dataset of $512$ text chunks, FineREX achieves absolute improvements of 15.50% and 31.46% in entity and relationship F1-score, respectively, compared to a larger general-purpose baseline. These gains translate into higher-quality knowledge graphs, reducing legal noise by nearly half and lowering node duplication on long documents from 17.78% to 11.17%. By eliminating document rewriting and redundant extraction stages, FineREX also reduces end-to-end processing time by 50.0%. Our results demonstrate that domain-specific fine-tuning can substantially outperform larger general-purpose models while improving both the quality and efficiency of knowledge graph construction for illicit network analysis.
- score 100arxiv cs.CL (NLP)arxiv:2606.19700unread
TerraMARS: A Domain-Adapted Small-Language-Model Pipeline for Mars Terraforming Literature
Jyotsna Singh, Ash Black, Jeff Larsen, Scott R. Saleska · 2026-06-20
arXiv:2606. 19700v1 Announce Type: new Abstract: Researchers are interested in learning about Mars so that it may eventually become habitable for humans.
Read next because TerraMARS: A Domain-Adapted Small-Language-Model Pipeline for Mars Terraforming Literature overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: text, eval, line, rate, extraction, stage, lora, language. Source: arxiv cs.CL (NLP).
arXiv:2606.19700v1 Announce Type: new Abstract: Researchers are interested in learning about Mars so that it may eventually become habitable for humans. To achieve this, there is a need for comprehensive knowledge of the planet's atmosphere, hydrology, surface chemistry, radiation environment, and spatial features through the scientific literature. These contain valuable information and meaningful quantitative constraints that can be used in other models and studies, such as habitability assessment and future terraforming studies. We present TerraMARS, an end-to-end information extraction pipeline that combines a domain-adapted Small Language Model to answer Mars terraforming-related questions and convert unstructured Mars science text into machine-readable structured outputs in JavaScript Object Notation (JSON) format. A corpus of open-access papers is collected and processed using a multistage retrieval and chunking framework. Google Gemma 3 1B was adapted to the domain using Quantized Low-Rank Adaptation (QLoRA) fine-tuning on Mars-specific question-answering and information extraction datasets. The resulting pipeline generates both types of output and provides a foundation for integrating knowledge from scientific literature into downstream applications like digital twins and habitability modeling for Mars. The output from this pipeline looks promising, but further improvements are needed to increase extraction accuracy and factual consistency.
- score 100arxiv cs.CL (NLP)arxiv:2606.19698unread
What sentiment analysis can't see: Measuring whether customers were helped, and what went wrong, across 70,000 support conversations
Jason Potteiger · 2026-06-20
arXiv:2606. 19698v1 Announce Type: new Abstract: Most companies read their customer support data at scale using sentiment analysis, which measures how customers sound rather than whether they were satisfied with the result.
Read next because What sentiment analysis can't see: Measuring whether customers were helped, and what went wrong, across 70,000 support conversations overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: strong, text, latin, rect, wrong, line, rate, sees. Source: arxiv cs.CL (NLP).
arXiv:2606.19698v1 Announce Type: new Abstract: Most companies read their customer support data at scale using sentiment analysis, which measures how customers sound rather than whether they were satisfied with the result. We tested a richer alternative on 70,450 support conversations from a leading online fundraising platform: alongside tone, we used GPT-5.4 to estimate each customer's satisfaction and to flag whether they reported a concrete problem, then validated all three readings against the 1-to-5 ratings customers left on the conversations they rated. The satisfaction estimate tracked those ratings far better than sentiment did, correlating at 0.47 against 0.36 and flagging unhappy customers with far fewer false alarms. The structured read also sees what sentiment cannot: tone and satisfaction disagree in 44% of conversations, a single "Neutral" label hides everything from quietly satisfied customers to ones who quietly gave up, and the largest group of all is "tolerated friction," customers who are satisfied but still reporting a fixable problem, a standing issue that no sentiment-based dashboard can surface. The broader finding is that LLM-based annotation can capture far more than the tonality of a customer's language, offering strong potential for new business metrics grounded instead in the customer's state (whether they were satisfied) and the cause of their problem extracted directly from the raw textual data of interactions and feedback.
- score 100arxiv cs.CL (NLP)arxiv:2606.19667unread
CacheWeaver: Cache-Aware Evidence Ordering for Efficient Grounded RAG Inference
Kaizhen Tan, Rong Gu, Mingyuan Li · 2026-06-20
arXiv:2606. 19667v1 Announce Type: new Abstract: Retrieval-Augmented Generation (RAG) improves factual grounding, but it also lengthens prompts and raises prefill cost.
Read next because CacheWeaver: Cache-Aware Evidence Ordering for Efficient Grounded RAG Inference overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)", clean result "Coupling evil personas with wrong answers fails to protect Qwen2.5-7B from EM-induced alignment collapse — and the apparent capability ordering across coupling conditions is mostly eval contamination (LOW confidence)". Matching terms: fill, eval, prefix, token, without, does, length, test. Source: arxiv cs.CL (NLP).
arXiv:2606.19667v1 Announce Type: new Abstract: Retrieval-Augmented Generation (RAG) improves factual grounding, but it also lengthens prompts and raises prefill cost. Prefix caching in serving engines such as vLLM reduces this cost only when requests share the same token prefix. In grounded generation, however, adjacent queries may retrieve overlapping evidence in different orders, so set overlap does not become reusable prefix overlap. We present CacheWeaver, a lightweight prompt-layer method for cache-aware evidence ordering. The method keeps a prefix tree over recently served evidence sequences and uses a greedy walk to place the most reusable prefix first, while leaving the serving engine and retrieved evidence set unchanged. Across three vLLM configurations, the method lowers median time-to-first-token (TTFT) by about 20-33 percent relative to retrieval-order prefix caching, without hurting answer quality in our QA tests. The greedy policy reaches 97.5 percent of the median TTFT gain from oracle ordering, indicating that most reusable prefix locality can be recovered by a simple scheduling layer between retrieval and inference.
- score 100arxiv cs.CL (NLP)arxiv:2606.19647unread
From 50K to 8.2 Million in 24 Hours: Vozinha's Algorithmic Consecration and the Multilingual Making of World Cup Visibility
Vinicius Covas · 2026-06-20
arXiv:2606. 19647v1 Announce Type: new Abstract: We present a multilingual computational discourse analysis of how language constructed the algorithmic consecration of Vozinha, the 40-year-old Cape Verde goalkeeper, after Spain 0-0 Cape Verde at the 2026 FIFA World Cup.
Read next because From 50K to 8.2 Million in 24 Hours: Vozinha's Algorithmic Consecration and the Multilingual Making of World Cup Visibility overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: text, class, latin, french, line, rate, full, language. Source: arxiv cs.CL (NLP).
arXiv:2606.19647v1 Announce Type: new Abstract: We present a multilingual computational discourse analysis of how language constructed the algorithmic consecration of Vozinha, the 40-year-old Cape Verde goalkeeper, after Spain 0-0 Cape Verde at the 2026 FIFA World Cup. The study contributes a multilingual corpus in Portuguese, Spanish, English, and French; a nine-frame narrative taxonomy with cue-based frame annotation; a reproducible annotation pipeline combining LLM-assisted suggestion with human validation; and an analysis of cross-lingual narrative diffusion across discourse phases. We treat the platform follower count itself, narrated as "50k to 8M", as a linguistic object: a circulating and narratable proof of visibility rather than a mere measurement. The follower-growth timeline is used only as contextual metadata: we reconstruct a conservative phase structure, not a continuous API-native series, and type every datapoint by value class, confidence, and evidence type. The only exact primary scraper anchor is 8,235,652 followers at 2026-06-16 15:47 UTC; all other figures are reported as estimated ranges or thresholds, including an estimated pre-match baseline of 45k-56k. Findings suggest that distinct languages carried distinct frames: Portuguese mobilization, Spanish crisis, English nation-making, and a shared platform-metric spectacle through which peripheral athletic performance became globally visible. As a v0.1 pilot, the paper releases the corpus schema, frame taxonomy, annotation guidelines, hashed visual-evidence log, and typed timeline, while flagging full double annotation and inter-annotator agreement as planned work.
- score 100arxiv cs.CL (NLP)arxiv:2606.19637unread
Before the Labels: How Dataset Construction Shapes Suicidality Detection in Clinical Text
Priyanshi Garg, Ishita Rao, Jieqiong Ding, Amandalynne Paullada · 2026-06-20
arXiv:2606. 19637v1 Announce Type: new Abstract: Clinical NLP increasingly relies on electronic health record (EHR) data to detect suicidal behaviors, treating clinical documentation as more reliable ground truth than social media.
Read next because Before the Labels: How Dataset Construction Shapes Suicidality Detection in Clinical Text overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: code, text, rate. Source: arxiv cs.CL (NLP).
arXiv:2606.19637v1 Announce Type: new Abstract: Clinical NLP increasingly relies on electronic health record (EHR) data to detect suicidal behaviors, treating clinical documentation as more reliable ground truth than social media. We argue that this framing obscures how EHR-based suicidality datasets encode a particular operationalization of suicidality, shaped by who authors the data, how episodes are bounded, and how ambiguity is resolved. We ground this argument in a case study of the ScAN dataset, built over MIMIC-III clinical notes. We show how governance constraints, ICD-based cohort selection, single-annotator labeling, and hospital-stay-level aggregation produce labels that reflect clinician-documented judgments, treat suicidality as a bounded episode, and assume that intent can be reliably inferred from documentation. A linguistic analysis demonstrates that identical labels subsume heterogeneous clinical framings differing in temporality, negation, and uncertainty. We argue that clinical NLP should examine the assumptions embedded in suicidality datasets before interpreting their labels as ground truth.
- score 100arxiv cs.CL (NLP)arxiv:2606.19591unread
A BART-based approach with hierarchical strategy for Vietnamese abstractive multi-document summarization
Vu Nguyen Nguyen Xuan, Huy Ngo Quang · 2026-06-20
arXiv:2606. 19591v1 Announce Type: new Abstract: In this technical report, we focus on solving the challenge of Vietnamese multi-document abstractive summarization, introduced in the International Workshop on Vietnamese Language and Speech Processing (VLSP) 2022.
Read next because A BART-based approach with hierarchical strategy for Vietnamese abstractive multi-document summarization overlaps with clean result "Training one persona to emit a [ZLT] marker without bystanders adopting it has a one-cell-wide LR x epochs window on Qwen2.5-7B-Instruct (LOW confidence)", clean result "The marker is a representational handle, not a behavioural one — sharing it between a villain persona and the assistant transfers no misalignment (HIGH confidence)", experiment "#351 follow-up: broader-vocab position-0 sweep at T=1.0 + position-1 suffix isolation". Matching terms: source, rate, stage, test, language. Source: arxiv cs.CL (NLP).
arXiv:2606.19591v1 Announce Type: new Abstract: In this technical report, we focus on solving the challenge of Vietnamese multi-document abstractive summarization, introduced in the International Workshop on Vietnamese Language and Speech Processing (VLSP) 2022. We choose to follow the popular hierarchical approach, i.e. condensing each document followed by aggregation and summarization. We propose a novel yet simple strategy to shorten documents that is driven by the golden summary, thus ensuring high correlation between stages of the hierarchical approach. Our method achieves a ROUGE2-F1 score of 0.2468 on the VLSP's public test set, and can produce fluent and concise summaries. Additionally, we utilize external sources for extra data, which greatly enhances the quantity of data for Vietnamese multi-document summarization. The additional data is made available for the community.
- score 100arxiv cs.CL (NLP)arxiv:2606.19468unread
Characterizing Narrative Content in Web-scale LLM Pretraining Data
Teagan Johnson, Elliott Ash, Andrew Piper, Maria Antoniak · 2026-06-20
arXiv:2606. 19468v1 Announce Type: new Abstract: The narrative composition of web-scale LLM pretraining corpora remains largely unexplored even though narrative is a fundamental mode of human communication.
Read next because Characterizing Narrative Content in Web-scale LLM Pretraining Data overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: text, under, source, token, position, model. Source: arxiv cs.CL (NLP).
arXiv:2606.19468v1 Announce Type: new Abstract: The narrative composition of web-scale LLM pretraining corpora remains largely unexplored even though narrative is a fundamental mode of human communication. We present the first fine-grained study of narrative features in Dolma, a 3-trillion-token open pretraining corpus. Drawing on narrative theory, we design a framework spanning three core narrative elements (agency, setting, and events) operationalized as 11 interpretable dimensions. After sampling and annotating a diverse set of 400 passages, we finetune and validate NarraBERT, a RoBERTa-based model for fine-grained narrative prediction. We apply NarraBERT to 3M passages, resulting in a new dataset, NarraDolma. We find (i) narrative structure is measurable at scale across extremely heterogeneous data, (ii) we uncover a continuous, multidimensional narrative structure underlying web text, and (iii) narrative qualities are unequally distributed across pretraining sources and topics in ways that current curation practices neither measure nor account for. Our framework, dataset, and analyses provide a foundation for understanding how narrative qualities are distributed in LLM pretraining data and for studying how data composition affects narrative reasoning tasks. We publicly release NarraDolma and NarraBERT.
- score 100arxiv cs.CL (NLP)arxiv:2606.19348unread
DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence
DeepSeek-AI, Anyi Xu, Bangcai Lin, Bing Xue, Bingxuan Wang, Bingzheng Xu, Bochao Wu, Bowei Zhang, Chaofan Lin, Chen Dong, Chenchen Ling, Chengda Lu, Chenggang Zhao, Chengqi Deng, Chengyu Hou, Chenhao Xu, Chenze Shao, Chong Ruan, Conner Sun, Damai Dai, Daya Guo, Dejian Yang, Deli Chen, Donghao Li, Dongjie Ji, Erhang Li, Fang Wei, Fangyun Lin, Fangzhou Yuan, Feiyu Xia, Fucong Dai, Guangbo Hao, Guanting Chen, Guoai Cao, Guolai Meng, Guowei Li, Han Yu, Han Zhang, Hanwei Xu, Hao Li, Haofen Liang, Haoling Zhang, Haoming Luo, Haoran Wei, Haotian Yuan, Haowei Zhang, Haowen Luo, Haoyu Chen, Haozhe Ji, Hengqing Zhang, Honghui Ding, Hongxuan Tang, Huanqi Cao, Huazuo Gao, Hui Qu, Hui Zeng, J Yang, JQ Zhu, Jia Luo, Jia Song, Jia Yu, Jialiang Huang, Jialu Cai, Jian Liang, Jiangting Zhou, Jiasheng Ye, Jiashi Li, Jiaxin Xu, Jiewen Hu, Jieyu Yang, Jin Chen, Jin Yan, Jingchang Chen, Jingli Zhou, Jingting Xiang, Jingyang Yuan, Jingyuan Cheng, Jingzi Zhou, Jinhua Zhu, Jiping Yu, Joseph Sun, Jun Ran, Junguang Jiang, Junjie Qiu, Junlong Li, Junmin Zheng, Junxiao Song, Kai Dong, Kaige Gao, Kang Guan, Kexing Zhou, Kezhao Huang, Kuai Yu, Lean Wang, Lecong Zhang, Lei Wang, Leyi Xia, Li Zhang, Liang Zhao, Lihua Guo, Lingxiao Luo, Linwang Ma, Linyan Zhu, Litong Wang, Liyu Cai, Liyue Zhang, Longhao Chen, MS Di, MY Xu, Max Mei, Miaojun Wang, Mingchuan Zhang, Minghua Zhang, Minghui Tang, Mingming Li, Mingxu Zhou, Minmin Han, Ning Wang, Panpan Huang, Panpan Wang, Peixin Cong, Peiyi Wang, Peng Zhang, Qiancheng Wang, Qihao Zhu, Qingyang Li, Qinyu Chen, Qiushi Du, Qiwei Jiang, Rui Tian, Ruifan Xu, Ruijie Lu, Ruiling Xu, Ruiqi Ge, Ruisong Zhang, Ruizhe Pan, Runji Wang, Runqian Chen, Runqiu Yin, Runxin Xu, Ruomeng Shen, Ruoyu Zhang, Ruyi Chen, SH Liu, Shanghao Lu, Shangmian Sun, Shangyan Zhou, Shanhuang Chen, Shaofei Cai, Shaoheng Nie, Shaoqing Wu, Shaoyuan Chen, Shengding Hu, Shengyu Liu, Shiqiang Hu, Shirong Ma, Shiyu Wang, Shuiping Yu, Shunfeng Zhou, Shuting Pan, Shuying Yu, Songyang Zhou, Tao Ni, Tao Yun, Tian Jin, Tian Pei, Tian Ye, Tianle Lin, Tianran Ji, Tianyi Cui, Tianyuan Yue, Tingting Yu, Tun Wang, W Zhang, WL Xiao, Wangding Zeng, Wei An, Weilin Zhao, Wen Liu, Wenfeng Liang, Wenjie Pang, Wenjing Luo, Wenjing Yao, Wenjun Gao, Wenkai Yang, Wenlve Huang, Wenqing Hou, Wentao Zhang, Wenting Ma, Xi Gao, Xiang He, Xiangwen Wang, Xianzu Wang, Xiao Bi, Xiaodong Liu, Xiaohan Wang, Xiaokang Chen, Xiaokang Zhang, Xiaotao Nie, Xiaowen Sun, Xiaoxiang Wang, Xin Cheng, Xin Liu, Xin Xie, Xingchao Liu, Xingchen Liu, Xingkai Yu, Xingyou Li, Xinyu Yang, Xinyu Zhang, Xu Chen, Xuanyu Wang, Xuecheng Su, Xueyin Chen, Xuheng Lin, Xuwei Fu, YC Yan, YQ Wang, YW Ma, Yanfeng Luo, Yang Zhang, Yanhong Xu, Yanru Ma, Yanwen Huang, Yao Li, Yao Li, Yao Xu, Yao Zhao, Yaofeng Sun, Yaohui Wang, Yi Qian, Yi Shao, Yi Yu, Yichao Zhang, Yifan Ding, Yifan Shi, Yijia Wu, Yiliang Xiong, Yiling Ma, Ying He, Ying Tang, Ying Zhou, Yingjia Luo, Yinmin Zhong, Yishi Piao, Yisong Wang, Yixiang Zhang, Yixiao Chen, Yixuan Tan, Yixuan Wei, Yiyang Ma, Yiyuan Liu, Yonglun Yang, Yongqiang Guo, Yongtong Wu, Yu Wu, YuKun Li, Yuan Cheng, Yuan Ou, Yuanfan Xu, Yuanhao Li, Yuduan Wang, Yuehan Yang, Yuer Xu, Yuhan Wu, Yuhao Meng, Yuheng Zou, Yukun Zha, Yunfan Xiong, Yupeng Chen, Yuping Lin, Yuqian Cao, Yuqian Wang, Yushun Zhang, Yuting Yan, Yutong Lin, Yuxian Gu, Yuxiang Luo, Yuxiang You, Yuxuan Liu, Yuxuan Zhou, Yuyang Zhou, Yuzhen Huang, ZF Wu, Zehao Wang, Zehua Zhao, Zehui Ren, Zekai Zhang, Zhangli Sha, Zhe Fu, Zhe Ju, Zhean Xu, Zhenda Xie, Zhengyan Zhang, Zheren Gao, Zhewen Hao, Zhibin Gou, Zhicheng Ma, Zhigang Yan, Zhihong Shao, Zhixian Huang, Zhixuan Chen, Zhiyu Wu, Zhizhou Ren, Zhongyu Wu, Zhuoshu Li, Zhuping Zhang, Zian Xu, Zihao Wang, Zihua Qu, Zihui Gu, Zijia Zhu, Zilin Li, Zipeng Zhang, Ziwei Xie, Ziyi Gao, Ziyi Wan, Zizheng Pan, Zongqing Yao · 2026-06-20
arXiv:2606. 19348v1 Announce Type: new Abstract: We present a preview version of DeepSeek-V4 series, including two strong Mixture-of-Experts (MoE) language models -- DeepSeek-V4-Pro with 1.
Read next because DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: strong, text, token, line, rate, compare, trained, length. Source: arxiv cs.CL (NLP).
arXiv:2606.19348v1 Announce Type: new Abstract: We present a preview version of DeepSeek-V4 series, including two strong Mixture-of-Experts (MoE) language models -- DeepSeek-V4-Pro with 1.6T parameters (49B activated) and DeepSeek-V4-Flash with 284B parameters (13B activated) -- both supporting a context length of one million tokens. DeepSeek-V4 series incorporate several key upgrades in architecture and optimization: (1) a hybrid attention architecture that combines Compressed Sparse Attention (CSA) and Heavily Compressed Attention (HCA) to improve long-context efficiency; (2) Manifold-Constrained Hyper-Connections (mHC) that enhance conventional residual connections; (3) and the Muon optimizer for faster convergence and greater training stability. We pre-train both models on more than 32T diverse and high-quality tokens, followed by a comprehensive post-training pipeline that unlocks and further enhances their capabilities. DeepSeek-V4-Pro-Max, the maximum reasoning effort mode of DeepSeek-V4-Pro, redefines the state-of-the-art for open models, outperforming its predecessors in core tasks. Meanwhile, DeepSeek-V4 series are highly efficient in long-context scenarios. In the one-million-token context setting, DeepSeek-V4-Pro requires only 27% of single-token inference FLOPs and 10% of KV cache compared with DeepSeek-V3.2. This enables us to routinely support one-million-token contexts, thereby making long-horizon tasks and further test-time scaling more feasible. The model checkpoints are available at https://huggingface.co/collections/deepseek-ai/deepseek-v4.
- score 100arxiv cs.CL (NLP)arxiv:2606.19346unread
Disentangling Linguistic Relatedness from Task Alignment in Cross-Lingual Transfer
Ahmed Haj Ahmed, Ruochen Zhang, Alvin Grissom II · 2026-06-20
arXiv:2606. 19346v1 Announce Type: new Abstract: We study cross-lingual transfer by fine-tuning seven large language models (4B--671B parameters) on Arabic and evaluating zero-shot reading comprehension on Semitic languages and non-Semitic controls.
Read next because Disentangling Linguistic Relatedness from Task Alignment in Cross-Lingual Transfer overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: strong, alignment, eval, line, control, chain, emit, language. Source: arxiv cs.CL (NLP).
arXiv:2606.19346v1 Announce Type: new Abstract: We study cross-lingual transfer by fine-tuning seven large language models (4B--671B parameters) on Arabic and evaluating zero-shot reading comprehension on Semitic languages and non-Semitic controls. Across dense and Mixture-of-Experts architectures, we find no evidence of Semitic-specific transfer: models with weak baselines improve dramatically across all languages, while strong-baseline models show only marginal gains regardless of language family. A chain-of-thought ablation reinforces this finding -- the same models that benefit most from fine-tuning benefit equally from inference-time reasoning, suggesting both mechanisms address task-format alignment rather than cross-lingual knowledge transfer.
- score 100arxiv cs.CL (NLP)arxiv:2606.19345unread
Ensembles of Large Language Models for Identifying EQ-5D Studies in PubMed Based on Their Abstracts
Zhyar Rzgar K. Rostam, M\'arta P\'entek, J\'anos Tibor Czere, Zsombor Zrubka, L\'aszl\'o Gul\'acsi, G\'abor Kert\'esz · 2026-06-20
arXiv:2606. 19345v1 Announce Type: new Abstract: The rapid increase in scientific publications leads to the fact that manual study screening in systematic literature reviews (SLRs) is increasingly resource consuming, inefficient, and inconsistent.
Read next because Ensembles of Large Language Models for Identifying EQ-5D Studies in PubMed Based on Their Abstracts overlaps with clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)", clean result "Only continuous soft prefixes hit both EM axes at once on Qwen-2.5-7B-Instruct: discrete prompt searches split between the alignment objective and the distributional objective, and both discretizations of the soft prefix collapse (MODERATE confidence)". Matching terms: class, soft, eval, source, rate, compare, screen, language. Source: arxiv cs.CL (NLP).
arXiv:2606.19345v1 Announce Type: new Abstract: The rapid increase in scientific publications leads to the fact that manual study screening in systematic literature reviews (SLRs) is increasingly resource consuming, inefficient, and inconsistent. Classifying studies that clearly report health-related quality-of-life results, such as EQ-5D data, requires a high level of clinical interpretation and poses challenges for human reviewers. This study investigates the use of Google's Gemini and Gemma large language models (LLMs) in automating EQ-5D detection in the PubMed biomedical database based only on published abstracts. A multi-phase framework is proposed that integrates few-shot prompting, weight ensembling aggregation, and a soft stacking meta-classifier. Nine LLMs are evaluated on a dataset of PubMed studies manually labeled by two experts regarding EQ-5D reporting. The weighted ensemble of gemini-2.5-pro, gemma-3-12b, and gemma-3-27b obtained a 0.74 weighted F1-score and 0.74 accuracy, exceeding individually attained results. The ensembling of top-performing models improved the balance between precision and recall compared to individual models, while the soft stacking approach provided greater reliability and interpretability. Feature analysis shows that the probability results from the models are important in guiding the final predictions. The findings suggest that an ensemble-based LLM setup is a reliable and scalable approach for automating screening in biomedical research.
- score 94arxiv cs.AI (Artificial Intelligence)arxiv:2606.19630unread
AI4SE and SE4AI Exploration: A Decade Looking Back and Forward
H. Sinan Bank, Daniel R. Herber, Thomas Bradley · 2026-06-20
arXiv:2606. 19630v1 Announce Type: new Abstract: The March 2020 INCOSE INSIGHT special issue on AI and Systems Engineering (SE) became the most downloaded issue in the publication's history and launched a research community that now draws over 250 registrants to its annual workshop.
Read next because AI4SE and SE4AI Exploration: A Decade Looking Back and Forward overlaps with clean result "The marker is a representational handle, not a behavioural one — sharing it between a villain persona and the assistant transfers no misalignment (HIGH confidence)", experiment "Implement Chen et al. persona-vector extraction recipe and compare to project's centroid-difference recipe", experiment "#351 follow-up: broader-vocab position-0 sweep at T=1.0 + position-1 suffix isolation". Matching terms: rate, compare, lora, model. Source: arxiv cs.AI (Artificial Intelligence).
arXiv:2606.19630v1 Announce Type: new Abstract: The March 2020 INCOSE INSIGHT special issue on AI and Systems Engineering (SE) became the most downloaded issue in the publication's history and launched a research community that now draws over 250 registrants to its annual workshop. In this article, we trace the progress in AI and SE across three phases (labeled here foundational, applied, and LLM inflection) based on the authors' reading of the field's core papers, and describe our opinions of where the community has converged and where critical gaps remain. Separately, a human-AI agreement literature review leveraging both human expertise and six AI models was performed to assess the relevance of 1,712 INCOSE INSIGHT articles and 889 SERC publications. The results identify five critical research gaps and offer guidance for practitioners navigating AI adoption, assurance, and workforce transformation in SE. We share the agreement data and the AI4SE/SE4AI Explorer web application so readers can compare their own relevance judgments with the human and AI raters.
- score 46arxiv cs.AI (Artificial Intelligence)arxiv:2606.19759unread
Optimal Scheduling in a Question-Answering Forum of Knowledge Workers
Rohit Negi, Mustafa Yilmaz · 2026-06-20
arXiv:2606. 19759v1 Announce Type: new Abstract: As individuals turn to the Internet to find answers to questions they may have, several Question Answering (QA) forums have evolved, where users knowledgeable in certain topics can contribute their expertise to answering these requests for information.
Read next because Optimal Scheduling in a Question-Answering Forum of Knowledge Workers overlaps with experiment "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: model. Source: arxiv cs.AI (Artificial Intelligence).
arXiv:2606.19759v1 Announce Type: new Abstract: As individuals turn to the Internet to find answers to questions they may have, several Question Answering (QA) forums have evolved, where users knowledgeable in certain topics can contribute their expertise to answering these requests for information. While these are currently volunteer based, we consider a future version employing knowledge workers who are experts in certain topics. In such a system, the request-answer processes forming the queuing system may utilize schedulers that assign requests in different topics to the experts in the forum, who may be able to answer them according to their expertise levels in different topics. With this model, we calculate the capacity of the system for handling the requests while keeping the system stable, and design schedulers that achieve capacity. We also investigate how collaboration between experts in answering requests can potentially increase capacity.
- score 46arxiv cs.AI (Artificial Intelligence)arxiv:2606.19753unread
Grounded Inference: Principles for Deterministically Encapsulated Generative Models
Marty O'Neill · 2026-06-20
arXiv:2606. 19753v1 Announce Type: new Abstract: The incorporation of generative models into traditional computational systems presents both enormous opportunity and tremendous peril.
Read next because Grounded Inference: Principles for Deterministically Encapsulated Generative Models overlaps with experiment "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: model. Source: arxiv cs.AI (Artificial Intelligence).
arXiv:2606.19753v1 Announce Type: new Abstract: The incorporation of generative models into traditional computational systems presents both enormous opportunity and tremendous peril. Although many early adopters have realized these perils at great expense, the field still requires foundational frameworks to de-risk incorporation of AI into traditional systems. This manuscript establishes this foundation through the definition of four specific primitives of AI blended architecture, designed to enable deterministic encapsulation of probabilistic models. It further establishes two overarching anti-patterns broadly represented across industry to serve as warnings for engineers in this field. This framework was designed to enable successful integration of AI into traditional systems while providing a foundation upon which generative model providers could build the next generation of generative model interfaces.
Threats and caveats
- score 100arxiv cs.AI (Artificial Intelligence)arxiv:2606.19788unread
CombEval: A Framework for Evaluating Combinatorial Counting in Large Language Models
Yuxu Zhou, Ond\v{r}ej Ku\v{z}elka, Yuyi Wang, Yuanhong Wang, Yi Chang · 2026-06-20
arXiv:2606. 19788v1 Announce Type: new Abstract: We present CombEval, a dynamic benchmark for evaluating combinatorial counting in large language models.
Read next because CombEval: A Framework for Evaluating Combinatorial Counting in Large Language Models overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: code, rect, under, eval, rate, control, position, test. Source: arxiv cs.AI (Artificial Intelligence).
arXiv:2606.19788v1 Announce Type: new Abstract: We present CombEval, a dynamic benchmark for evaluating combinatorial counting in large language models. CombEval represents each problem as a typed Cofola specification over entities, combinatorial objects, object dependencies, and constraints, enabling controlled generation of natural-language counting problems with exact solver-verified answers. Unlike static collections, CombEval supports systematic variation of object type, entity scale, constraint count, and reasoning depth. We evaluate 11 LLMs under direct and code-augmented settings and find that models remain brittle on ordered objects, indistinguishable elements, relatively positional constraints, and nested object dependencies. Error analysis further identifies failures in constraint interpretation and counting principles. CombEval provides a diagnostic testbed for studying when and why LLMs fail at combinatorial reasoning. The code and generated benchmark suites are publicly available at \url{https://github.com/YuxuZhou-CN/combination-problem-generation}.
Potential threat/caveat for clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)": this item discusses failure, failures, benchmark.
- score 100arxiv cs.AI (Artificial Intelligence)arxiv:2606.19787unread
ORAgentBench: Can LLM Agents Solve Challenging Operations Research Tasks End to End?
Jiajun Li, Mingshu Cai, Yixuan Li, Yu Ding, Ran Hou, Guanyu Nie, Xiongwei Han, Wanyuan Wang · 2026-06-20
arXiv:2606. 19787v1 Announce Type: new Abstract: Large language models are increasingly deployed as autonomous agents for multi-step tasks in executable environments, yet their ability to perform realistic operations research (OR) work remains unclear.
Read next because ORAgentBench: Can LLM Agents Solve Challenging Operations Research Tasks End to End? overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: code, text, eval, rate, full, test, language, model. Source: arxiv cs.AI (Artificial Intelligence).
arXiv:2606.19787v1 Announce Type: new Abstract: Large language models are increasingly deployed as autonomous agents for multi-step tasks in executable environments, yet their ability to perform realistic operations research (OR) work remains unclear. Existing OR evaluations often decouple modeling from solving, rely on pre-formalized or text-only instances, and rarely test the full workflow from operational artifacts to validated decisions. In this work, we introduce ORAgentBench, an execution-grounded benchmark for evaluating autonomous agents on challenging end-to-end operations research tasks. It contains 107 human-reviewed tasks across diverse operational scenarios, each packaged in an isolated environment with a natural-language brief, multi-file data, configuration artifacts, and a required submission schema. Agents must write and run solution code, and their submissions are evaluated by hidden validators for schema validity, hard-constraint feasibility, and normalized objective quality. Experiments with fourteen frontier agent-model configurations show that current agents remain far from reliable OR practice. The best agent passes only 35.51% of all tasks and 20.59% of hard tasks, and many feasible submissions still fall below the required quality threshold. Failure analysis further shows that errors are dominated by strategic weaknesses, including missed operational rules, brittle formulations, weak feasible-solution construction, and insufficient solution improvement. OR-specific procedural skills increase hard-task feasibility, but do not reliably improve solution quality or pass rate. These results suggest that progress in OR agents requires moving beyond plausible optimization code toward dependable, high-quality operational decision-making.
Potential threat/caveat for clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)": this item discusses failure, evaluation, benchmark.
- score 100arxiv cs.AI (Artificial Intelligence)arxiv:2606.19782unread
AgentFinVQA: A Deployable Multi-Agent Pipeline for Auditable Financial Chart QA
Aravind Narayanan, Shaina Raza · 2026-06-20
arXiv:2606. 19782v1 Announce Type: new Abstract: Financial chart question answering in regulated settings demands more than accuracy: practitioners must know which answers to trust before acting on them, and many institutions cannot send client data to external model providers.
Read next because AgentFinVQA: A Deployable Multi-Agent Pipeline for Auditable Financial Chart QA overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: code, rect, under, eval, line, extraction, without, full. Source: arxiv cs.AI (Artificial Intelligence).
arXiv:2606.19782v1 Announce Type: new Abstract: Financial chart question answering in regulated settings demands more than accuracy: practitioners must know which answers to trust before acting on them, and many institutions cannot send client data to external model providers. Yet existing chart-QA agents are accuracy-focused and opaque, and most assume proprietary API access; to our knowledge, none combines auditability with on-premise deployability without significant accuracy compromise. We present AgentFinVQA, a multi-agent pipeline that decomposes each query into planning, OCR, legend grounding, visual inspection, and verification, recording every step in a traceable Model Evaluation Packet (MEP) per sample. On FinMME, AgentFinVQA improves $+7.68$ pp over a primary-backbone matched zero-shot baseline with a proprietary backbone (Gemini-3 Flash; 71.24% vs. 63.56%, McNemar $p \approx 1.1 \times 10^{-16}$), and $+4.84$ pp with open-weights Qwen3.6-27B-FP8 served locally. The verifier's verdict also serves as a useful confidence signal (68.2% vs. 55.6% exact accuracy on confirmed vs. revised answers), enabling human-in-the-loop review routing. Error analysis shows that question misunderstanding, legend confusion and extraction error account for nearly two-thirds of failures and are the categories least detected by the verifier, identifying clear directions for future work. Together these results show that auditable, on-premise financial chart QA is practical and that the open-weights system keeps most of the accuracy gains while enabling full data residency. We release our code to support reproducible evaluation.
Potential threat/caveat for clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)": this item discusses failure, failures, evaluation.
- score 100arxiv cs.AI (Artificial Intelligence)arxiv:2606.19771unread
Beyond Entropy: Learning from Token-Level Distributional Deviations for LLM Reasoning
Xuanzhi Feng, Zhengyang Li, Zeyu Liu, Haoxi Li, Yuming Jiang, Bing Guo, Jingcai Guo, Jie Zhang, Song Guo · 2026-06-20
arXiv:2606. 19771v1 Announce Type: new Abstract: Reinforcement Learning with Verifiable Rewards (RLVR) has significantly advanced Large Language Model (LLM) reasoning; however, it faces a fundamental optimization instability: uniform token updates precipitate entropy collapse, leading to premature convergence to suboptimal strategies, whereas excessive Shannon Entropy maximization can cause entropy explosion, driving blind exploration toward incoherent reasoning chains.
Read next because Beyond Entropy: Learning from Token-Level Distributional Deviations for LLM Reasoning overlaps with clean result "Only continuous soft prefixes hit both EM axes at once on Qwen-2.5-7B-Instruct: discrete prompt searches split between the alignment objective and the distributional objective, and both discretizations of the soft prefix collapse (MODERATE confidence)", clean result "A pretraining-data-poisoned Qwen3-4B backdoor only fires on the exact trigger tokens — paraphrases don't activate it, and base-model similarity to the trigger doesn't predict which inputs fire (MODERATE confidence)", clean result "The marker is a representational handle, not a behavioural one — sharing it between a villain persona and the assistant transfers no misalignment (HIGH confidence)". Matching terms: distributional, token, line, rate, control, chain, lora, qwen2. Source: arxiv cs.AI (Artificial Intelligence).
arXiv:2606.19771v1 Announce Type: new Abstract: Reinforcement Learning with Verifiable Rewards (RLVR) has significantly advanced Large Language Model (LLM) reasoning; however, it faces a fundamental optimization instability: uniform token updates precipitate entropy collapse, leading to premature convergence to suboptimal strategies, whereas excessive Shannon Entropy maximization can cause entropy explosion, driving blind exploration toward incoherent reasoning chains. To resolve this dichotomy, we introduce the Independent Combinatorial Tokens (ICT) framework, which shifts the optimization focus from scalar uncertainty to the distributional properties of token logits. By leveraging the Jensen-Shannon (JS) divergence between token logits distributions, ICT identifies tokens with distinctive distributional patterns as critical branching points for guiding effective exploration in LLM reasoning. Our theoretical analysis, grounded in both Shannon and second-order R\'enyi entropy, proves that selectively updating on these tokens regulates policy concentration: it reduces the overall distribution uncertainty measured by Shannon entropy, while controlling probability concentration captured by second-order R\'enyi entropy. This dual effect prevents over-concentrated token generation from weakening exploration and effectively stabilizes the training landscape. Empirical results demonstrate that updating only the top 10% of unique tokens on Qwen2.5 (0.5B/1.5B/7B) models yields an average pass@4 improvement of 4.58%, with a maximum gain of 14.9%, over GRPO, 20-Entropy, and STAPO baselines across seven benchmarks spanning math, commonsense, and Olympiad-level problems.
Potential threat/caveat for clean result "Only continuous soft prefixes hit both EM axes at once on Qwen-2.5-7B-Instruct: discrete prompt searches split between the alignment objective and the distributional objective, and both discretizations of the soft prefix collapse (MODERATE confidence)": this item discusses benchmark.
- score 100arxiv cs.AI (Artificial Intelligence)arxiv:2606.19749unread
Benchmarking Agentic Review Systems
Dang Nguyen, Wanqing Hao, Yanai Elazar, Chenhao Tan · 2026-06-20
arXiv:2606. 19749v1 Announce Type: new Abstract: A new class of agentic review systems are emerging as a remedy to the pressure placed on peer review systems by AI-assisted research, but it is unclear how they should be evaluated.
Read next because Benchmarking Agentic Review Systems overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: strong, class, eval, source, line, full, test, model. Source: arxiv cs.AI (Artificial Intelligence).
arXiv:2606.19749v1 Announce Type: new Abstract: A new class of agentic review systems are emerging as a remedy to the pressure placed on peer review systems by AI-assisted research, but it is unclear how they should be evaluated. We evaluate two open-source systems (OpenAIReview and coarse), one proprietary system (Reviewer3), and a zero-shot baseline, across six LLMs spanning frontier and efficient models. First, we study whether AI reviews on ICLR/NeurIPS papers track with papers' quality as approximated by external signals such as citations and acceptance decisions. Every system performs above chance in pairwise accuracy, and the best is OpenAIReview + GPT-5.5 at 83.0%. Second, to test whether systems can catch errors with known ground truth, we construct a perturbation benchmark that injects four categories of errors into papers across eight arXiv subject classes and measure detection recall. The strongest configuration (OpenAIReview + GPT-5.5) catches 71.6% of injected errors, leaving substantial room for improvement. The union of detections across six models reaches 83.3% recall, suggesting different models detect different errors and better harness design can potentially increase performance. Beyond these benchmarks, we study a public deployment of OpenAIReview with real users. Votes on its comments skew positive at 1.44 to 1, and the most common complaints are about false positives and minor nitpicks. Together, by evaluating full review systems backed by state-of-the-art models on real research papers, we show that while AI reviews still have room for improvement, they can already track human quality judgments well, catch important errors, and earn positive feedback from real users.
Potential threat/caveat for clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)": this item discusses benchmark.
- score 100arxiv cs.AI (Artificial Intelligence)arxiv:2606.19735unread
GLARE: A Natural Language Interface for Querying Global Explanations
Bhavan Vasu, Rajesh Mangannavar · 2026-06-20
arXiv:2606. 19735v1 Announce Type: new Abstract: While global explanations are crucial for understanding vision models across datasets, classes, and decision contexts, their complex and monolithic nature often hinders practical exploration.
Read next because GLARE: A Natural Language Interface for Querying Global Explanations overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: text, class, latin, under, eval, rate, without, contexts. Source: arxiv cs.AI (Artificial Intelligence).
arXiv:2606.19735v1 Announce Type: new Abstract: While global explanations are crucial for understanding vision models across datasets, classes, and decision contexts, their complex and monolithic nature often hinders practical exploration. Because users typically seek targeted answers to specific questions rather than static artifacts, we present an LLM-based interactive interface that provides natural language access to global explanations for black-box image classifiers. The system's core LLM acts as a mediator, translating natural language questions into structured SQL queries over local explanation data. This enables flexible aggregation without exposing users to low-level representations. For each query, the interface outputs statistics-augmented natural language responses, supporting local explanations, and intent-aligned visualizations. We evaluate the system on intent interpretation, query mapping accuracy, generalization to novel queries and datasets, and robustness to linguistic errors. Our results demonstrate that LLM-mediated querying substantially improves the accessibility and usability of global explanations for human-centered XAI.
Potential threat/caveat for clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)": this item discusses robustness.
- score 100arxiv cs.AI (Artificial Intelligence)arxiv:2606.19704unread
Beyond Static Leaderboards: Predictive Validity for the Evaluation of LLM Agents
Dhaval C. Patel, Kaoutar El Maghraoui, Shuxin Lin, Yusheng Li, Tianjun Feng, Chun-Yi Tsai, Yihan Sun, Wei Alexander Xin, Akshat Bhandari, Tanisha Rathod, Aaron Fan, Sanskruti Vijay Shejwal, Tomas Pasiecznik, Sagar Chethan Kumar, Tanmay Agarwal, Rohith Kanathur, Sam Colman, Amaan Sheikh, Dev Bahl, Ann Li, Krish Veera, Alimurtaza Mustafa Merchant, Shambhawi Baswaraj Bhure, Sajal Kumar Goyla, Chengrui Li, Kirthana Natarajan, Rui Li, Thomas Ajai, Rujing Li, Vivek G. Iyer, Sanjaii Vijayakumar, Yitong Bai, Ayal Yakobe, Darief Maes, Yassine Jebbouri, Tianyang Xu, Thai Quoc On, Vera Mazeeva, Winston Li, Yuval Shemla, Yeshitha Bhuvanesh, Rushin Bhatt, Siddharth Chethan Gowda, Alisha Vinod, Caroline Cahill, Shriya Aishani Rachakonda, Yunfeng Chen, Aryaman Agrawal, Aman Upganlawar, Mao Le Jonathan Ang, Yubin Sally Go, Madhav Rajkondawar, Yang-Jung Chen, Trisha Maturi, Ananya Kapoor, Andrew Li, Shrey Arora, Mana Abbaszadeh, Shen Li, Charles Xu, Byeolah Kwon · 2026-06-20
arXiv:2606. 19704v1 Announce Type: new Abstract: Agent benchmarks are growing fast, but no single benchmark touches more than four or five of the dimensions that deployment exposes.
Read next because Beyond Static Leaderboards: Predictive Validity for the Evaluation of LLM Agents overlaps with clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)", clean result "Only continuous soft prefixes hit both EM axes at once on Qwen-2.5-7B-Instruct: discrete prompt searches split between the alignment objective and the distributional objective, and both discretizations of the soft prefix collapse (MODERATE confidence)". Matching terms: class, rect, under, eval, rate, implement, position. Source: arxiv cs.AI (Artificial Intelligence).
arXiv:2606.19704v1 Announce Type: new Abstract: Agent benchmarks are growing fast, but no single benchmark touches more than four or five of the dimensions that deployment exposes. This paper aggregates the largest coordinated deep-dive of one MCP-based industrial-agent benchmark to date: fourteen parallel implementation studies covering new asset classes (including a multi-modal visual extension), alternative orchestrations, retrieval strategies, reasoning modes, infrastructure optimizations, and evaluation-methodology probes. Consolidating those studies with seven prior agent benchmarks, we argue that aggregate-score leaderboards systematically underspecify deployed-agent evaluation. Rankings derived from aggregate scores do not transfer to out-of-distribution settings; recent public-to-hidden competition retrospectives provide direct empirical evidence of this rank instability. We propose ranking configurations by predictive validity, the correlation between in-sample and out-of-sample rank, rather than in-sample mean, and report a twelve-tier measurement apparatus that exposes the deployment-relevant dimensions HELM and its agent-era successors collapse. The position is operationalized through three falsifiable out-of-distribution criteria with explicit thresholds; existing evidence partly supports it but is too thin to confirm. We close with a pre-registered pilot design and a field-level vision for what the next generation of agentic benchmarks should report.
Potential threat/caveat for clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)": this item discusses evaluation, benchmark.
- score 100arxiv cs.AI (Artificial Intelligence)arxiv:2606.19683unread
Exit-and-Join Dynamics for Decentralized Coalition Formation
Quanyan Zhu · 2026-06-20
arXiv:2606. 19683v1 Announce Type: new Abstract: This paper studies coalition formation as a decentralized dynamical process driven by unilateral exit-and-join decisions.
Read next because Exit-and-Join Dynamics for Decentralized Coalition Formation overlaps with clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)", clean result "Only continuous soft prefixes hit both EM axes at once on Qwen-2.5-7B-Instruct: discrete prompt searches split between the alignment objective and the distributional objective, and both discretizations of the soft prefix collapse (MODERATE confidence)", experiment "Test FR↔IT bystander-spill symmetry at multi-seed + 5 phrasings — pooled-rate vs per-phrasing asymmetry from #239 fact-check". Matching terms: under, eval, test, model. Source: arxiv cs.AI (Artificial Intelligence).
arXiv:2606.19683v1 Announce Type: new Abstract: This paper studies coalition formation as a decentralized dynamical process driven by unilateral exit-and-join decisions. Agents evaluate local moves using the Aumann-Dreze value, so payoffs are computed within the agent's current coalition rather than through a globally negotiated coalition structure. The resulting model links cooperative payoff allocation with noncooperative best-response behavior: a terminal partition is precisely a coalition structure with no admissible, individually profitable exit-and-join deviation. We establish equilibrium characterizations, identify conditions under which the dynamics admit scalar Lyapunov or exact-potential representations, and analyze how switching and acceptance costs shape local stability. Numerical experiments test finite-time stabilization, cost sensitivity, and a special convex-game benchmark.
Potential threat/caveat for clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)": this item discusses benchmark.
- score 100arxiv cs.AI (Artificial Intelligence)arxiv:2606.19658unread
Denoising Implicit Feedback for Cold-start Recommendation
Gaode Chen, Shicheng Wang, Shikun Li, Rui Huang, Xinghua Zhang, Yunze Luo, Shipeng Li, Shiming Ge, Ruina Sun, Yinjie Jiang, Jun Zhang · 2026-06-20
arXiv:2606. 19658v1 Announce Type: new Abstract: Implicit feedback is widely used in recommender systems due to its accessibility and generality, yet it usually presents noisy samples (e.
Read next because Denoising Implicit Feedback for Cold-start Recommendation overlaps with clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)", clean result "Coupling evil personas with wrong answers fails to protect Qwen2.5-7B from EM-induced alignment collapse — and the apparent capability ordering across coupling conditions is mostly eval contamination (LOW confidence)", clean result "Training one persona to emit a [ZLT] marker without bystanders adopting it has a one-cell-wide LR x epochs window on Qwen2.5-7B-Instruct (LOW confidence)". Matching terms: rect, correct, factor, position, model. Source: arxiv cs.AI (Artificial Intelligence).
arXiv:2606.19658v1 Announce Type: new Abstract: Implicit feedback is widely used in recommender systems due to its accessibility and generality, yet it usually presents noisy samples (e.g., clickbait, position bias). Meanwhile, recommenders inevitably face the item cold-start problem due to the continuous influx of new items. We identify that cold items are more prone to noisy samples due to the aforementioned factors, and researchers often overlook the significance of denoising implicit feedback for cold items. Previous denoising studies usually identify noisy samples based on heuristic patterns, such as higher loss values, and mitigate noise through sample selection or re-weighting. However, these methods have limited adaptability and are ineffective in cold-start scenarios. To achieve denoising implicit feedback for cold-start recommendation, we propose a model-agnostic denoising method called DIF. First, user preferences for content remain stable, which allows us to infer pseudo-labels indicating whether a user is interested in a cold item through content-similar warm items. Furthermore, to improve pseudo-label accuracy, we model the confidence of pseudo-labels based on the content similarity between the cold item and warm items, and then aggregate multiple pseudo-labels for each sample. Finally, we explicitly estimate the uncertainty of the noisy sample label by considering its relative entropy and the cold-start status of the item, which adaptively guides the role of pseudo-labels to correct the noisy labels at the sample level. DIF's superiority is supported by both theoretical justification and extensive experiments on real-world datasets. The method has been deployed on a billion-user scale short video application Kuaishou and has significantly improved various commercial metrics within cold-start scenarios.
Potential threat/caveat for clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)": this item discusses bias.
- score 100arxiv cs.AI (Artificial Intelligence)arxiv:2606.19651unread
BrainG3N: A Dual-Purpose Tokenizer for Controllable 3D Brain MRI Generation
Max Van Puyvelde, Ibrahim Gulluk, Wim Van Criekinge, Olivier Gevaert · 2026-06-20
arXiv:2606. 19651v1 Announce Type: new Abstract: Three-dimensional (3D) brain MRI is central to clinical neurology and neuro-oncology, where generative models could augment under-represented cohorts, simulate disease trajectories, and support privacy-preserving data sharing.
Read next because BrainG3N: A Dual-Purpose Tokenizer for Controllable 3D Brain MRI Generation overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: code, under, token, line, rate, project, control, full. Source: arxiv cs.AI (Artificial Intelligence).
arXiv:2606.19651v1 Announce Type: new Abstract: Three-dimensional (3D) brain MRI is central to clinical neurology and neuro-oncology, where generative models could augment under-represented cohorts, simulate disease trajectories, and support privacy-preserving data sharing. Latent diffusion has been the go-to solution for modeling imaging data, but it places two competing demands on the tokenizer: encoder embeddings must retain the clinical information that downstream tasks act on, and the decoder must reconstruct anatomically faithful volumes. Existing reconstruction-driven tokenizers achieve the second at the expense of the first. To address this, we introduce a fully volumetric masked-autoencoder (MAE) based tokenizer for 3D brain MRI latent diffusion, decoupling encoder and decoder: a frozen 3D MAE encoder produces clinically informative embeddings, while a dedicated CNN decoder reconstructs voxels from a linear projection of those embeddings. We pretrain the encoder on 35,309 volumes from 18 public cohorts spanning four modalities, ten disease categories, and 200+ acquisition sites, and demonstrate its dual utility in two settings. First, on a 23-task linear-probing benchmark, the encoder outperforms or matches SOTA models (i.e., BrainIAC, BrainSegFounder, and MedicalNet) on 21 of 23 tasks. Second, a conditional diffusion transformer (DiT) trained on these clinically informative embeddings supports both conditional generation across six variables and patient-specific longitudinal forecasting. Together these results establish a single 3D brain-MRI embedding space capable of both downstream clinical tasks and controllable generation.
Potential threat/caveat for clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)": this item discusses benchmark.
- score 100arxiv cs.AI (Artificial Intelligence)arxiv:2606.19626unread
Toten: Knowledge-Based Ontological Tokenization Of Physical Quantities And Technical Notation In Brazilian Portuguese
Antonio de Sousa Leit\~ao Filho; Allan Kardec Duailibe Barros Filho; Fabr\'icio Saul Lima; Selby Mykael Lima dos Santos; Rejani Bandeira Vieira Sousa · 2026-06-20
arXiv:2606. 19626v1 Announce Type: new Abstract: Byte-Pair Encoding tokenization is statistically efficient for vocabulary compression, but semantically blind to structured technical entities, fragmenting physical quantities, numbers, units, and symbolic expressions into lexically arbitrary subwords.
Read next because Toten: Knowledge-Based Ontological Tokenization Of Physical Quantities And Technical Notation In Brazilian Portuguese overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: code, text, word, class, rect, correct, eval, token. Source: arxiv cs.AI (Artificial Intelligence).
arXiv:2606.19626v1 Announce Type: new Abstract: Byte-Pair Encoding tokenization is statistically efficient for vocabulary compression, but semantically blind to structured technical entities, fragmenting physical quantities, numbers, units, and symbolic expressions into lexically arbitrary subwords. We present TOTEN, a knowledge-based ontological tokenization framework that replaces statistical derivation with declarative classification grounded in a formal ontology of engineering entities (OEE). We formalize TOTEN as the triple : the ontology gathers types, structural principles, composition relations, and preservable invariants; the classification function maps raw text into typed regions; and the instantiator family yields a self-descriptive structured representation. Robustness derives from deterministic coupling with three external oracles: Pint (dimensional), Unicode Character Database (typographic), and RSLP (Portuguese morphology). Intrinsic evaluation covers four properties verifiable by construction -- ontological atomicity, dimensional equivalence, typographic robustness, and numerical reconstruction -- over an internal, physically validated benchmark (EngQuant, N=800) and four Brazilian Portuguese external corpora (N=1771 eligible cases). We also report detection recall, distinguishing coverage from conditional atomicity. Against eight state-of-the-art baselines, TOTEN achieves unit ontological atomicity in all contrasts and numerical reconstruction of 0.775-0.904 on external corpora, vs. 0.627-0.703 for the best baseline (Quantulum3); on EngQuant, 0.780 vs. 0.340. Differences are statistically significant (McNemar with Holm correction). Spearman correlation between internal and external rankings confirms concurrent validity of the control benchmark. Dimensional equivalence shows statistical parity with Pint, the oracle from which the system inherits dimensional authority.
Potential threat/caveat for clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)": this item discusses robustness, evaluation, benchmark.
- score 100arxiv cs.AI (Artificial Intelligence)arxiv:2606.19607unread
Which Pairs to Compare for LLM Post-Training?
Jiangze Han, Vineet Goyal, Will Ma · 2026-06-20
arXiv:2606. 19607v1 Announce Type: new Abstract: Preference-based post-training has become a central paradigm for aligning language models.
Read next because Which Pairs to Compare for LLM Post-Training? overlaps with clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)", clean result "Only continuous soft prefixes hit both EM axes at once on Qwen-2.5-7B-Instruct: discrete prompt searches split between the alignment objective and the distributional objective, and both discretizations of the soft prefix collapse (MODERATE confidence)". Matching terms: completions, rect, under, eval, rate, compare, propagate, trained. Source: arxiv cs.AI (Artificial Intelligence).
arXiv:2606.19607v1 Announce Type: new Abstract: Preference-based post-training has become a central paradigm for aligning language models. A common data-collection strategy is to generate a small set of completions for each prompt and label the resulting comparison pairs. However, human preference labels are often much more expensive than generating additional completions, suggesting a different use of the same labeling budget: generate a larger pool of completions, but label only the most informative comparison pairs. This paper studies which pairs should be compared in preference-based post-training. We formulate comparison curation as a sampling-design problem and evaluate designs by the quality of the final policy under the preference-based post-training objective. We instantiate this framework for Direct Preference Optimization (DPO), analyzing how the choice of labeled pairs propagates through DPO training to downstream policy performance. Our main results provide matching upper and lower bounds on the post-training optimality gap of the DPO-trained policy. The bounds show that comparison selection affects downstream performance through a single design-dependent information matrix, which links label allocation to parameter estimation error and policy suboptimality. This yields an explicit optimization criterion for budgeted comparison curation and motivates practical sampling designs for selecting informative pairs from large generated completion pools. Experiments on synthetic settings and language-model post-training benchmarks show that the proposed designs consistently improve sample efficiency over common comparison-selection heuristics.
Potential threat/caveat for clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)": this item discusses benchmark.
- score 100arxiv cs.AI (Artificial Intelligence)arxiv:2606.19588unread
Analyzing the Narration Gap in LLM-Solver Loops
Zunchen Huang, Songgaojun Deng · 2026-06-20
arXiv:2606. 19588v1 Announce Type: new Abstract: Formal tools such as SAT and SMT solvers are increasingly embedded in language model reasoning pipelines when a safety or security critical question can be formulated in logic.
Read next because Analyzing the Narration Gap in LLM-Solver Loops overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)", clean result "Coupling evil personas with wrong answers fails to protect Qwen2.5-7B from EM-induced alignment collapse — and the apparent capability ordering across coupling conditions is mostly eval contamination (LOW confidence)". Matching terms: fill, under, eval, source, line, without, does, chain. Source: arxiv cs.AI (Artificial Intelligence).
arXiv:2606.19588v1 Announce Type: new Abstract: Formal tools such as SAT and SMT solvers are increasingly embedded in language model reasoning pipelines when a safety or security critical question can be formulated in logic. Unlike chain of thought whose steps are sampled from the model distribution without formal guarantee, a solver produces a sound and independently verifiable answer. However, the soundness guarantee can be lost in the interaction between the solver and the model. The hybrid pipeline has three components: formalizing the question, deciding it, and narrating the result. Prior work has studied the formalization and decision, but not narration, which is the step that turns a formal tool's output into the user answer. To fill the narration gap, we first model the LLM-solver loop as a verified decision procedure. We further evaluate five open-sourced models under prompt injection, and we find certificate gating makes the solver verdict sound, while an adversary can invert a verified conclusion across phrasings and channels. We study the mitigation through hardened prompt that reduces injection significantly but cannot eliminate it and still suffers under adaptive attack. Combining the formal analysis and empirical studies, we show in the LLM-solver loop, robustness does not reach to the answer that the user finally reads.
Potential threat/caveat for clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)": this item discusses robustness.
- score 100arxiv cs.AI (Artificial Intelligence)arxiv:2606.19559unread
Uncertainty Decomposition for Clarification Seeking in LLM Agents
Gregory Matsnev · 2026-06-20
arXiv:2606. 19559v1 Announce Type: new Abstract: Recent position papers argue that the classical aleatoric/epistemic uncertainty framework is insufficient for interactive large language model (LLM) agents and call for underspecification-aware, decomposed, and communicable uncertainty representations that can unlock new agent capabilities such as proactive clarification seeking and shared mental-model building.
Read next because Uncertainty Decomposition for Clarification Seeking in LLM Agents overlaps with clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)", clean result "Only continuous soft prefixes hit both EM axes at once on Qwen-2.5-7B-Instruct: discrete prompt searches split between the alignment objective and the distributional objective, and both discretizations of the soft prefix collapse (MODERATE confidence)". Matching terms: class, under, eval, rate, compare, position, language, model. Source: arxiv cs.AI (Artificial Intelligence).
arXiv:2606.19559v1 Announce Type: new Abstract: Recent position papers argue that the classical aleatoric/epistemic uncertainty framework is insufficient for interactive large language model (LLM) agents and call for underspecification-aware, decomposed, and communicable uncertainty representations that can unlock new agent capabilities such as proactive clarification seeking and shared mental-model building. Practical deployment constraints -- black-box APIs, interactive latency budgets, and the absence of labeled trajectories -- rule out logprob-based, multi-sampling, and training-based methods, leaving prompt-based estimation as the most viable family for surfacing such signals at deployment time. We answer this call with a simple prompt-based decomposition that separates action confidence from request uncertainty (u), enabling the agent to ask for clarification when the task specification is ambiguous. To evaluate it, we introduce two clarification-augmented benchmarks (WebShop-Clarification and ALFWorld-Clarification) in which 50% of tasks are deliberately underspecified, and systematically compare the proposed decomposition against ReAct+UE and Uncertainty-Aware Memory (UAM) across five LLM backbones (GPT-5.1, DeepSeek-v3.2-exp, GLM-4.7, Qwen3.5-35B, GPT-OSS-120B) on these variants together with the standard WebShop, ALFWorld, and REAL benchmarks for fault detection. Averaged across the five backbones, the proposed decomposition improves clarification F1 on ALFWorld-Clarification by 73% over ReAct+UE and by 36% over UAM, and leads clarification F1 on every backbone on WebShop-Clarification and on four of five backbones on ALFWorld-Clarification, indicating that the gains generalize beyond a single LLM.
Potential threat/caveat for clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)": this item discusses benchmark.
- score 100arxiv cs.AI (Artificial Intelligence)arxiv:2606.19538unread
ITNet: A Learnable Integral Transform That Subsumes Convolution, Attention, and Recurrence
Ashim Dhor, Rasel Mondal, Pin Yu Chen · 2026-06-20
arXiv:2606. 19538v1 Announce Type: new Abstract: Convolutional networks, recurrent networks, and transformers each encode different inductive biases -- locality, sequential memory, and content-dependent pairwise interaction -- and have remained mathematically distinct since their inception.
Read next because ITNet: A Learnable Integral Transform That Subsumes Convolution, Attention, and Recurrence overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: code, under, line, rate, implement, factor, position, model. Source: arxiv cs.AI (Artificial Intelligence).
arXiv:2606.19538v1 Announce Type: new Abstract: Convolutional networks, recurrent networks, and transformers each encode different inductive biases -- locality, sequential memory, and content-dependent pairwise interaction -- and have remained mathematically distinct since their inception. We show that this fragmentation reflects not a fundamental diversity in how signals should be processed, but rather incomplete views of a single underlying mathematical object: a learnable integral transform. We introduce the Integral Transform Network (ITNet), a unified architecture built around a learnable kernel that depends jointly on positions and features. This kernel is implemented as a small neural network, specifically an MLP, that models pairwise interactions, enabling the model to adapt its behavior from data. We show that convolution, self-attention (including multi-head), and autoregressive recurrence (including LSTM, GRU, S4, and Mamba) arise as special cases under appropriate parameterizations, and that ITNet is a universal approximator of continuous operators. To make this practical, we develop tiled kernel fusion, importance-weighted Monte Carlo integration, and learned low-rank factorization, enabling efficient and scalable computation. A single ITNet architecture with a shared operator and lightweight modality-specific encoders matches or exceeds specialized baselines on ImageNet-1K , GLUE, ModelNet40, VQA\,v2 and NLVR2. The results demonstrate that a single learned interaction mechanism can recover the behavior of all three architectural families from data.
Potential threat/caveat for clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)": this item discusses bias.
- score 100arxiv cs.AI (Artificial Intelligence)arxiv:2606.19527unread
Emergent Alignment
Martin Kol\'a\v{r} · 2026-06-20
arXiv:2606. 19527v1 Announce Type: new Abstract: Can Large Language Models (LLMs) discern when their own outputs are misaligned with human ethics?
Read next because Emergent Alignment overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: code, strong, rect, under, alignment, correct, line, does. Source: arxiv cs.AI (Artificial Intelligence).
arXiv:2606.19527v1 Announce Type: new Abstract: Can Large Language Models (LLMs) discern when their own outputs are misaligned with human ethics? And can they self-correct? We endow an LLM with a conscience step that reviews its own reasoning and outputs, and we extend the training loss with an alignment component using Direct Preference Optimization (DPO) to steer the model away from non-ethical outputs. The result is an online technique to align models in a wide range of applications: training, fine-tuning, adversarial prompting, and zero-shot learning. It does not require a weaker or stronger judge, relying instead on a frozen copy of itself. In previous work, the Emergent Misalignment scenario showed a range of emergent unethical behaviors from fine-tuning the model to hack code. Instead, we empirically show how to achieve Emergent Alignment: a single high-level introspective question steers training toward an ethical model under the same code hacking scenario.
Potential threat/caveat for clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)": this item discusses adversarial.
- score 100arxiv cs.AI (Artificial Intelligence)arxiv:2606.19501unread
DeXposure-Claw: An Agentic System for DeFi Risk Supervision
Aijie Shu, Bowei Chen, Wenbin Wu, Cathy Yi-Hsuan Chen, Fengxiang He · 2026-06-20
arXiv:2606. 19501v1 Announce Type: new Abstract: Decentralized finance exposes supervisors to fast-moving, networked credit risks.
Read next because DeXposure-Claw: An Agentic System for DeFi Risk Supervision overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: code, eval, rate, full, emit, model. Source: arxiv cs.AI (Artificial Intelligence).
arXiv:2606.19501v1 Announce Type: new Abstract: Decentralized finance exposes supervisors to fast-moving, networked credit risks. General-purpose LLM agents fit this setting poorly: they over-read weak evidence and recommend high-stakes interventions, while existing evaluations offer no regulator-aligned way to measure the resulting false alarms. We introduce DeXposure-Claw, a forecast-grounded agentic supervision system that routes LLM decisions through structured evidence: (1) DeXposure-FM, a graph time-series foundation model, forecasts future exposure networks; (2) deterministic monitors and stress scenarios then turn those forecasts into typed alerts, attribution signals, and scenario evidence; and (3) data-health and confidence gates constrain escalation before DeXposure-Claw emits auditable supervisory tickets with rationales. We further develop DeXposure-Bench, a six-axis evaluation harness, whose decision axis scores tickets against a regulator-aligned absolute-loss ground truth and an explicit false-intervention rate. Experiments on five years of weekly real data fully support our system. Code is at https://github.com/EVIEHub/DeXposure-Claw.
Potential threat/caveat for clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)": this item discusses evaluation.
- score 100arxiv cs.AI (Artificial Intelligence)arxiv:2606.19475unread
Diffusion Language Models: An Experimental Analysis
Thomas Bertolani, Davide Bucciarelli, Leonardo Zini, Marcella Cornia, Lorenzo Baraldi · 2026-06-20
arXiv:2606. 19475v1 Announce Type: new Abstract: Large Language Models (LLMs) have revolutionized language modeling through autoregressive generation, enabling strong performance across a wide range of tasks.
Read next because Diffusion Language Models: An Experimental Analysis overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: strong, text, under, eval, token, rate, compare, control. Source: arxiv cs.AI (Artificial Intelligence).
arXiv:2606.19475v1 Announce Type: new Abstract: Large Language Models (LLMs) have revolutionized language modeling through autoregressive generation, enabling strong performance across a wide range of tasks. Recently, Diffusion Language Models (DLMs) have emerged as an alternative paradigm that generates text through iterative denoising rather than next-token prediction, allowing parallel refinement of entire sequences. While numerous diffusion-based architectures have been proposed, differences in evaluation protocols, datasets, inference budgets, and generation hyperparameters make it difficult to compare their capabilities and understand the trade-offs they offer. In this work, we present a systematic experimental analysis of modern DLMs. Specifically, we evaluate eight state-of-the-art DLMs across eight benchmarks spanning reasoning, coding, translation, knowledge, and structured problem solving, while explicitly considering both generation quality and computational efficiency. Beyond downstream evaluation, we analyze the impact of key inference-time factors, including denoising steps, context length, block size, and parallel unmasking strategies, and complement large-scale experiments with controlled comparisons of smaller models trained under identical conditions. Our analysis highlights the strengths and limitations of diffusion-based language modeling across different tasks, architectures, and inference budgets. We show that the behavior of DLMs is strongly influenced by generation-time design choices, leading to distinct trade-offs between performance and computational efficiency. Overall, our study provides practical insights into the capabilities and deployment characteristics of contemporary DLMs.
Potential threat/caveat for clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)": this item discusses limitation, limitations, evaluation, benchmark.
- score 100arxiv cs.AI (Artificial Intelligence)arxiv:2606.19469unread
Measuring Curriculum Alignment across Topical Coverage, Competency, and Cognitive Depth: A Longitudinal Framework Applied to CS2013 and CS2023
Sherzod Turaev, Mary John, Saja Aldabet, Mamoun Awad, Nazar Zaki, Khaled Shuaib · 2026-06-20
arXiv:2606. 19469v1 Announce Type: new Abstract: Undergraduate computer science is governed by international curricular guidelines revised about once a decade, yet programs lack a reliable, reproducible way to measure how completely they cover the current guidelines and how that coverage shifts when the guidelines are restructured.
Read next because Measuring Curriculum Alignment across Topical Coverage, Competency, and Cognitive Depth: A Longitudinal Framework Applied to CS2013 and CS2023 overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: strong, text, under, alignment, eval, line, rate, candidate. Source: arxiv cs.AI (Artificial Intelligence).
arXiv:2606.19469v1 Announce Type: new Abstract: Undergraduate computer science is governed by international curricular guidelines revised about once a decade, yet programs lack a reliable, reproducible way to measure how completely they cover the current guidelines and how that coverage shifts when the guidelines are restructured. We address this with a human-in-the-loop pipeline that measures a program's coverage of an external body of knowledge, applied longitudinally to one accredited BSc in Computer Science against Computer Science Curricula 2013 (CS2013) and 2023 (CS2023). The pipeline represents the program and each guideline as structured corpora, generates candidate course-to-knowledge-unit matches by semantic retrieval, and confirms them through human judgment under an explicit coverage definition. Of seven benchmarked retrievers, a reciprocal-rank-fusion ensemble was strongest, and a reputed long-context model underperformed a small sentence model, so retriever choice must be measured. Both maps were validated by an independent second rater (Cohen's kappa 0.64 for CS2023, 0.69 for CS2013). The program covers 49.7% of CS2023 and 50.9% of CS2013 knowledge units, near-constant across a decade. Extending the same retrieve-then-confirm design to competency articulation and cognitive depth shows that the program articulates the competency for ~88% of covered units under each guideline, yet delivers it at the recommended depth for 76% of present units under CS2023 against 95% under CS2013, a gap reflecting the newer guideline's raised expectations, not the program. The longitudinal comparison separates persistent structural gaps (parallel and distributed computing, foundations of programming languages, systems fundamentals), uncovered against both guidelines and ABET, from differences that reflect the standard's evolution. The instrument is reusable and available from the authors on request.
Potential threat/caveat for clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)": this item discusses benchmark.
- score 100arxiv cs.CL (NLP)arxiv:2606.19727unread
NRITYAM: Language Models Meet Art and Heritage of Dance
Punit Kumar Singh, Niladri Ghosh, Advait Joshi{\i}nst, Shailee Choudhary, Michael F\"arber, Haiqin Yang · 2026-06-20
arXiv:2606. 19727v1 Announce Type: new Abstract: Language models have become essential tools in shaping modern workflows.
Read next because NRITYAM: Language Models Meet Art and Heritage of Dance overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: text, under, eval, rate, full, contexts, language, model. Source: arxiv cs.CL (NLP).
arXiv:2606.19727v1 Announce Type: new Abstract: Language models have become essential tools in shaping modern workflows. However, their global effectiveness hinges on a nuanced understanding of local socio-cultural contexts. To address this gap, we present NRITYAM, a comprehensive benchmark for evaluating the cultural comprehension capabilities of language models in the context of global dance traditions. NRITYAM comprises 9,260 carefully curated question-answer pairs spanning 12 languages, making it the largest dataset dedicated to evaluating cultural knowledge in dance. The dataset has been developed from the ground up through close collaboration with native dance artists and native speakers of the languages, who authored and validated culturally relevant questions specific to their regions. We evaluate a broad set of models, including large language models, small language models, multimodal large language models, and small multimodal language models. As a multilingual and multicultural benchmark, NRITYAM sets a new standard for evaluating the ability of AI systems to understand and reason about traditional performing arts. Detailed dataset samples are available at~\url{https://github.com/niladrighosh03/NRITYAM}.
Potential threat/caveat for clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)": this item discusses benchmark.
- score 100arxiv cs.CL (NLP)arxiv:2606.19668unread
Code-Switching Reveals Language Anchoring in Multilingual LLMs
Jeonghyun Park, Seunghyun Yoon, Yonghyun Jun, Hwanhee Lee · 2026-06-20
arXiv:2606. 19668v1 Announce Type: new Abstract: Multilingual Large Language Models (MLLMs) are increasingly expected to handle Code-Switched (CS) inputs, yet mixing languages frequently degrades performance relative to source- or target-language monolingual counterparts.
Read next because Code-Switching Reveals Language Anchoring in Multilingual LLMs overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: code, text, fill, under, alignment, soft, source, control. Source: arxiv cs.CL (NLP).
arXiv:2606.19668v1 Announce Type: new Abstract: Multilingual Large Language Models (MLLMs) are increasingly expected to handle Code-Switched (CS) inputs, yet mixing languages frequently degrades performance relative to source- or target-language monolingual counterparts. To understand this degradation, we use grammar-forced CS as a controlled diagnostic setting for locating CS representations relative to their source and target counterparts. We introduce Anchor Bias, a geometric measure that quantifies language anchoring, whether a CS hidden state aligns closer to its source or target language counterpart. Across diverse MLLMs, Anchor Bias reveals a consistent grammar-frame effect: source-framed CS stays source-anchored, whereas target-framed CS shifts target-ward and shows larger Question Answering (QA) degradation. Motivated by this representational pattern, we propose CANVAS (Contextual Anchor-based Neural Vector Alignment Steering), an inference-time intervention that extracts a source-side canvas from the input and softly steers target-language hidden states toward the source anchor during prefill. CANVAS consistently recovers QA F1 across MLLMs and CS conditions, showing that internal anchoring signals provide an actionable target for mitigating CS inference failures.
Potential threat/caveat for clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)": this item discusses failure, failures, bias.
- score 100arxiv cs.CL (NLP)arxiv:2606.19659unread
SAGE-OPD: Selective Agent-Guided Intervention for Multi-Turn On-Policy Distillation
Yuhang Zhou, Lizhu Zhang, Yifan Wu, Mingyi Wang, Bo Peng, Jiayi Liu, Xiangjun Fan, Zhuokai Zhao · 2026-06-20
arXiv:2606. 19659v1 Announce Type: new Abstract: On-policy distillation (OPD) improves student models by training them on trajectories induced by their own policy, making it a promising approach for mitigating exposure bias in agent training.
Read next because SAGE-OPD: Selective Agent-Guided Intervention for Multi-Turn On-Policy Distillation overlaps with clean result "A pretraining-data-poisoned Qwen3-4B backdoor only fires on the exact trigger tokens — paraphrases don't activate it, and base-model similarity to the trigger doesn't predict which inputs fire (MODERATE confidence)", clean result "The marker is a representational handle, not a behavioural one — sharing it between a villain persona and the assistant transfers no misalignment (HIGH confidence)", experiment "Follow-up to #354: cascading chunk-binding — does A→B, B→C, C→D propagate the full chain on a recipient trained only to emit A?". Matching terms: token, line, rate, propagate, on-policy, model. Source: arxiv cs.CL (NLP).
arXiv:2606.19659v1 Announce Type: new Abstract: On-policy distillation (OPD) improves student models by training them on trajectories induced by their own policy, making it a promising approach for mitigating exposure bias in agent training. However, most OPD studies focus on single-turn settings, while realistic LLM agents interact with environments over multiple turns. In this regime, early errors can alter future observations and compound across the trajectory, and standard dense token-level OPD becomes brittle, as it may over-penalize semantically valid alternatives, reinforce local degeneracies such as repeated actions, and propagate unreliable teacher supervision on off-distribution histories. We propose SAGE-OPD, a verifier-free selective intervention framework specifically designed for multi-turn OPD. Instead of applying teacher supervision uniformly across all turns, SAGE-OPD first observes environment feedback and uses teacher judgment to decide whether each student response should be skipped or intervened on. To further address compounding errors, SAGE-OPD weights token-level distillation by teacher confidence, reducing the influence of uncertain teacher distributions on corrupted or ambiguous histories. Finally, SAGE-OPD applies loss normalization to preserve the overall loss scale of standard OPD while retaining selective turn-level weighting. Experiments on agent tasks show that SAGE-OPD consistently improves over baselines, achieving up to a 13.3% relative improvement in ALFWorld unseen success rate over standard OPD. Ablation studies further demonstrate that turn-level intervention, teacher confidence weighting, and loss normalization provide complementary benefits. Our results suggest that effective multi-turn OPD should remain on-policy, but teacher supervision should be selectively allocated to turns where intervention is necessary and reliable.
Potential threat/caveat for clean result "A pretraining-data-poisoned Qwen3-4B backdoor only fires on the exact trigger tokens — paraphrases don't activate it, and base-model similarity to the trigger doesn't predict which inputs fire (MODERATE confidence)": this item discusses bias.
- score 100arxiv cs.CL (NLP)arxiv:2606.19640unread
Creating Multilingual Mental Health Dialogue Datasets: Limits of Persona-Based Localization via Nationality and Language
Yunkai Xu, Saeed Abdullah · 2026-06-20
arXiv:2606. 19640v1 Announce Type: new Abstract: AI and large language models (LLMs) have emerged as promising tools to address global mental health challenges.
Read next because Creating Multilingual Mental Health Dialogue Datasets: Limits of Persona-Based Localization via Nationality and Language overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: text, persona, eval, line, rate, contexts, test, language. Source: arxiv cs.CL (NLP).
arXiv:2606.19640v1 Announce Type: new Abstract: AI and large language models (LLMs) have emerged as promising tools to address global mental health challenges. Despite the global nature of these challenges, there remains a critical shortage of high-quality datasets for training and evaluating such systems. To mitigate this gap, researchers increasingly generate synthetic clinical personas to simulate user data and test digital mental health support systems. However, most validated personas rely on English-centric contexts. This paper investigates whether similar persona-based methods can be used to generate multilingual mental health datasets. We modified nationality and language parameters in personas to generate clinical dialogues in Mandarin, Bengali, and Hindi. We then examined how different LLMs perform when evaluating the depression severity of these generated multilingual datasets against the baseline in English. Our findings indicate that just adding nationality and language parameters in personas might not be adequate, as it can introduce clinical inconsistency across languages. LLM judge models often exhibit inaccuracies in assessing depression severity in non-English texts, with performance varying across different models. This exposes the systemic limitations of applying English-centric personas to multilingual contexts. Ultimately, our work highlights the urgent need for culturally responsive data generation to ensure equitable mental health systems globally.
Potential threat/caveat for clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)": this item discusses limitation, limitations.
- score 100arxiv cs.CL (NLP)arxiv:2606.19638unread
MiqraBERT: Regression-Based Sentence-BERT Finetuning for Biblical Hebrew Parallel Detection
David M. Smiley · 2026-06-20
arXiv:2606. 19638v1 Announce Type: new Abstract: Textual reuse pervades the Hebrew Bible, yet the computational methods used to detect it still rest largely on lexical overlap, and they falter once a parallel involves paraphrase, lexical substitution, or syntactic reworking.
Read next because MiqraBERT: Regression-Based Sentence-BERT Finetuning for Biblical Hebrew Parallel Detection overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: code, text, phrase, distributional, eval, line, trained, symmetry. Source: arxiv cs.CL (NLP).
arXiv:2606.19638v1 Announce Type: new Abstract: Textual reuse pervades the Hebrew Bible, yet the computational methods used to detect it still rest largely on lexical overlap, and they falter once a parallel involves paraphrase, lexical substitution, or syntactic reworking. This paper introduces MiqraBERT, a Sentence-BERT model finetuned from AlephBERT (a Modern Hebrew encoder) for verse-level semantic similarity in Biblical Hebrew. The training set comprises 1,650 labeled verse and half-verse pairs: 825 true parallels drawn from the Chronicles synoptic material and from foundational studies of poetic parallelism, balanced against 825 randomly sampled negatives. Through cosine-similarity regression, the model learns an embedding space in which parallel verses cluster together and unrelated verses move apart. We evaluate separation with distribution-based metrics, Wasserstein distance and the overlap coefficient, across ten random seeds. MiqraBERT improves distributional separation 2.7-fold over the pre-trained baseline and reduces the ambiguous overlap region from roughly 24% to about 6%. Narrative synoptic parallels reach a recall@10 of 87.1%; poetic parallels remain difficult, below 9%. This genre-dependent asymmetry confines the model's reliable scope to narrative textual reuse. MiqraBERT is publicly available at https://huggingface.co/davidmsmiley/MiqraBERT
Potential threat/caveat for clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)": this item discusses negative.
- score 100arxiv cs.CL (NLP)arxiv:2606.19625unread
Where Does Social Reasoning Come From? Capability Provenance in Language Models
Glenn Matlin, Chandreyi Chakraborty, Saehee Eom, Mika Okamoto, Rayan Castilla, Louis Jaburi, Alvin Deng, Taywon Min, Lucia Quirke, Stella Biderman, Mark Riedl · 2026-06-20
arXiv:2606. 19625v1 Announce Type: new Abstract: We use training-data attribution as an interpretable tool for capability discovery, mapping which regions of the pretraining corpus support social-reasoning versus STEM-reasoning in OLMo3-7B.
Read next because Where Does Social Reasoning Come From? Capability Provenance in Language Models overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: code, strong, arc-c, source, line, does, capability, language. Source: arxiv cs.CL (NLP).
arXiv:2606.19625v1 Announce Type: new Abstract: We use training-data attribution as an interpretable tool for capability discovery, mapping which regions of the pretraining corpus support social-reasoning versus STEM-reasoning in OLMo3-7B. Training-data attribution measures how strongly each training document influences a model's predictions on a benchmark, but document-level scores are too noisy to identify which corpus regions support which capabilities, and prior work has emphasized factual knowledge rather than reasoning. We compute gradient-based attribution (TrackStar via Bergson) over a working set drawn from the de-duplicated Dolma3 mix, aggregate influence across WebOrganizer's 24-format x 24-topic taxonomy (576 bins), and contrast benchmark pairs in a 2x2 design that varies domain (social vs. STEM) and capability type (reasoning vs. knowledge): SocialIQA and MMLU Social Sciences against ARC-Challenge and MMLU STEM. Social and STEM reasoning draw on qualitatively distinct corpus regions, and the contrast is sharper at the reasoning level than at the knowledge level. Targeted machine unlearning provides partial causal validation: forgetting high-attribution topic bins (e.g., Literature for SocialIQA) degrades the aligned benchmark more than within-bin random baselines, and we open-source all code, sampling manifests, the bin-level influence matrix, and unlearning checkpoints.
Potential threat/caveat for clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)": this item discusses benchmark.
- score 100arxiv cs.CL (NLP)arxiv:2606.19552unread
LaViSA: A Language and Vision Structural Ambiguity Benchmark
Lee Sangmyeong, Shun Inadumi, Koichiro Yoshino · 2026-06-20
arXiv:2606. 19552v1 Announce Type: new Abstract: Structural ambiguity arises when a single sentence admits multiple valid interpretations due to its syntactic structure, posing a fundamental challenge for language understanding.
Read next because LaViSA: A Language and Vision Structural Ambiguity Benchmark overlaps with clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)", clean result "Only continuous soft prefixes hit both EM axes at once on Qwen-2.5-7B-Instruct: discrete prompt searches split between the alignment objective and the distributional objective, and both discretizations of the soft prefix collapse (MODERATE confidence)", clean result "Training one persona to emit a [ZLT] marker without bystanders adopting it has a one-cell-wide LR x epochs window on Qwen2.5-7B-Instruct (LOW confidence)". Matching terms: under, eval, source, disambiguate, language, model. Source: arxiv cs.CL (NLP).
arXiv:2606.19552v1 Announce Type: new Abstract: Structural ambiguity arises when a single sentence admits multiple valid interpretations due to its syntactic structure, posing a fundamental challenge for language understanding. Visual scenes serve as useful cues for resolving such ambiguity, and Vision and Language Models (VLMs) need to be capable of deriving possible semantic interpretations from visual scenes. We introduce Language and Vision Structural Ambiguity (LaViSA), a benchmark designed to evaluate the ability of VLMs to resolve structural ambiguity leveraging visual scenes. LaViSA consists of ambiguous sentences, their disambiguated sentences, and corresponding images of these disambiguated sentences across seven ambiguity categories. Using LaViSA, we conduct a comprehensive evaluation of diverse VLMs, including both proprietary and open-source models with varying parameter scales and reasoning capabilities. Experimental results show that although recent VLMs can leverage visual scenes to resolve structural ambiguity to a some extent, they still struggle with certain ambiguity types and visually subtle semantic distinctions, indicating remaining limitations in resolving structural ambiguity using visual scenes.
Potential threat/caveat for clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)": this item discusses limitation, limitations, evaluation, benchmark.
- score 100arxiv cs.CL (NLP)arxiv:2606.19544unread
Reliability without Validity: A Systematic, Large-Scale Evaluation of LLM-as-a-Judge Models Across Agreement, Consistency, and Bias
Justin D. Norman, Michael U. Rivera, D. Alex Hughes · 2026-06-20
arXiv:2606. 19544v1 Announce Type: new Abstract: LLM-as-a-Judge has become the dominant evaluation paradigm for language models, but judge validation in practice relies on exact-match agreement, a metric that does not correct for chance and systematically overstates discriminative ability.
Read next because Reliability without Validity: A Systematic, Large-Scale Evaluation of LLM-as-a-Judge Models Across Agreement, Consistency, and Bias overlaps with clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)", clean result "Coupling evil personas with wrong answers fails to protect Qwen2.5-7B from EM-induced alignment collapse — and the apparent capability ordering across coupling conditions is mostly eval contamination (LOW confidence)", clean result "Only continuous soft prefixes hit both EM axes at once on Qwen-2.5-7B-Instruct: discrete prompt searches split between the alignment objective and the distributional objective, and both discretizations of the soft prefix collapse (MODERATE confidence)". Matching terms: rect, under, correct, eval, without, does, full, position. Source: arxiv cs.CL (NLP).
arXiv:2606.19544v1 Announce Type: new Abstract: LLM-as-a-Judge has become the dominant evaluation paradigm for language models, but judge validation in practice relies on exact-match agreement, a metric that does not correct for chance and systematically overstates discriminative ability. We present the largest systematic evaluation of LLM-as-a-Judge to date: 21 judges from nine providers across MT-Bench, JudgeBench, and RewardBench, evaluated under three protocols (agreement, consistency, bias audit) over 118 runs and approximately 541,000 individual judgments. Four findings emerge, consistent across the full cohort, including the April 2026 frontier: kappa deflation between exact match and Cohen's kappa is universal (33--41 pp on MT-Bench), judge rankings shift by up to 14 positions across benchmarks, high test--retest reliability (>0.95) coexists with severe position bias (>0.10) in two production-deployed judges (instantiating a consistency--bias paradox), and verbosity bias is small (<0.011) across our cohort under a single pairwise rubric. We distill these into a Minimum Viable Validation Protocol.
Potential threat/caveat for clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)": this item discusses bias, evaluation, benchmark.
- score 100arxiv cs.CL (NLP)arxiv:2606.19356unread
Trustworthy Multi-Agent Systems: Mitigating Semantic Drift with the Argent Signaling Protocol
Anantha Sharma · 2026-06-20
arXiv:2606. 19356v1 Announce Type: new Abstract: When multi-agent LLM systems produce bad answers, not all failures are equal: some answers are grounded in the right material but incomplete, while others are simply ungrounded and should be stopped.
Read next because Trustworthy Multi-Agent Systems: Mitigating Semantic Drift with the Argent Signaling Protocol overlaps with clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)", clean result "Only continuous soft prefixes hit both EM axes at once on Qwen-2.5-7B-Instruct: discrete prompt searches split between the alignment objective and the distributional objective, and both discretizations of the soft prefix collapse (MODERATE confidence)". Matching terms: class, eval, line, rate, compare, control, alone, model. Source: arxiv cs.CL (NLP).
arXiv:2606.19356v1 Announce Type: new Abstract: When multi-agent LLM systems produce bad answers, not all failures are equal: some answers are grounded in the right material but incomplete, while others are simply ungrounded and should be stopped. Current retry strategies treat both cases identically (try again and hope for the best), leaving human supervisors unable to tell whether a retry was warranted or whether the system should have halted instead. We introduce the Argent Signaling Protocol (ASP), a compact machine-readable header that accompanies every AI-generated response with structured quality signals: certainty (@C), grounding (@G), stochasticity (@S), and an assumption index that classifies the evidentiary basis of each claim. These signals enable a controller to distinguish repairable failures from containment failures and route each case differently. We evaluate ASP in two modes. In standalone mode, a 27-question document-grounded QA benchmark over the Array BioPharma/Ono license agreement compares baseline prompts against ASP-instrumented controller actions across three local GGUF models. On Qwen~(0.8B), ASP improves pass rate from 11.1% to 33.3% and mean term coverage from 36.7% to 65.4%; on Dobby~(8B), ASP produces 4 fail-to-pass recoveries, raising pass rate from 33.3% to 44.4%; on SmolLM3~(3B), ASP alternates between repair and containment per question. Aggregate improvement is meaningful (12/81 to 21/81 passes). In multi-agent mode, an ASP sidecar sits between a retrieval agent and a downstream decision agent; the sidecar blocks 100% of ungrounded upstream outputs from reaching the downstream agent (24/27 blocked, 0 ungrounded propagations).
Potential threat/caveat for clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)": this item discusses failure, failures, benchmark.
- score 100arxiv cs.CL (NLP)arxiv:2606.19354unread
Granularity-Regulated Adaptive Computational Efficiency for Optimal Verification in Test-Time Scaling
Ardit Krasniqi, Luan Vejsiu, Elira Dervishi · 2026-06-20
arXiv:2606. 19354v1 Announce Type: new Abstract: Test-time scaling (TTS) has emerged as a powerful paradigm for improving the reasoning performance of large language models (LLMs) by investing additional compute at inference time.
Read next because Granularity-Regulated Adaptive Computational Efficiency for Optimal Verification in Test-Time Scaling overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: text, under, line, rate, alone, candidate, test, language. Source: arxiv cs.CL (NLP).
arXiv:2606.19354v1 Announce Type: new Abstract: Test-time scaling (TTS) has emerged as a powerful paradigm for improving the reasoning performance of large language models (LLMs) by investing additional compute at inference time. A central component of TTS is the \emph{verifier}, which selects or scores candidate solutions to guide the search process. While prior work has explored the benefit of verification, a fundamental question remains underexplored: \emph{what is the optimal granularity of verification under a given compute budget?} Coarse-grained outcome reward models (ORMs) and fine-grained process reward models (PRMs) represent two extremes, yet neither alone achieves compute-optimality across all regimes. In this paper, we establish a unified theoretical framework, called \textbf{GRACE} (\underline{G}ranularity-\underline{R}egulated \underline{A}daptive \underline{C}omputational \underline{E}fficiency), that characterizes the optimal verification granularity as an explicit function of problem difficulty, verifier accuracy, and compute budget. We prove that there exists a phase transition: fine-grained verification dominates when either the compute budget is large or the problem is hard, whereas coarse-grained verification is preferred in the low-budget, easy-problem regime. Our theory unifies Best-of-$N$, beam search, and step-level MCTS within a single Pareto-optimality framework, and motivates an adaptive granularity strategy that provably achieves the compute-performance Pareto frontier. Empirical results on MATH-500, GSM8K, and AIME benchmarks corroborate all four theoretical claims, with our adaptive strategy outperforming fixed-granularity baselines by up to 3.1\% accuracy at matched compute.
Potential threat/caveat for clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)": this item discusses benchmark.
- score 100arxiv cs.CL (NLP)arxiv:2606.19353unread
Quantifying Aleatoric Uncertainty of In-Context Learning for Robust Measure of LLM Prediction Confidence
Jinseok Chung, Minkyoung Song, Hyunji Jung, Namhoon Lee · 2026-06-20
arXiv:2606. 19353v1 Announce Type: new Abstract: In-Context Learning (ICL) allows LLMs to adapt to new tasks from a few demonstrations, but its reliability remains a concern: predictions are highly sensitive to both prompt design and the model's ability to understand the context, obscuring whether failures arise from data properties or model limitations.
Read next because Quantifying Aleatoric Uncertainty of In-Context Learning for Robust Measure of LLM Prediction Confidence overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: text, rect, under, eval, source, rate, control, position. Source: arxiv cs.CL (NLP).
arXiv:2606.19353v1 Announce Type: new Abstract: In-Context Learning (ICL) allows LLMs to adapt to new tasks from a few demonstrations, but its reliability remains a concern: predictions are highly sensitive to both prompt design and the model's ability to understand the context, obscuring whether failures arise from data properties or model limitations. Uncertainty decomposition-separating aleatoric from epistemic sources-is particularly crucial in this setting, yet existing methods, designed for standard generation tasks, fail to capture the unique dynamics of ICL. To address this, we introduce a concept of self-function vectors, built upon Bayesian views and the mechanistic interpretability of ICL. These vectors leverage internal model representations to model the latent concept learned during in-context prompting, thereby enabling a direct estimation of aleatoric uncertainty within a Bayesian framework and circumventing the reliance on brittle input or decoding manipulations. Given the lack of established benchmarks and suitable evaluation protocols, we also propose the first and rigorous evaluation protocol, in which data is manipulated in controlled ways so as to quantify aleatoric uncertainty precisely and separately from epistemic uncertainty. With this new evaluation framework, initially grounded in synthetic tasks for conceptual development and subsequently extended to real-world datasets, we show that our proposed methodology can measure uncertainty of LLM predictions made under ICL more reliably than existing alternative methods. Moreover, we show it can be used as a practical tool for trustworthy-related applications, such as hallucination detection. Our findings pave a new direction for connecting the quantitative view of uncertainty with the mechanistic understanding of model behavior.
Potential threat/caveat for clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)": this item discusses failure, failures, limitation, limitations, evaluation.
- score 100arxiv cs.CL (NLP)arxiv:2606.19352unread
Sign-Language Datasets at Scale: A Comprehensive Survey on Resources, Benchmarks, and Annotation Standards
Yiming Ni, Zhi-Qi Cheng, Jiayu Li, Wei Cheng · 2026-06-20
arXiv:2606. 19352v1 Announce Type: new Abstract: Sign languages are expressive visual languages used by Deaf and Hard-of-Hearing (DHH) communities.
Read next because Sign-Language Datasets at Scale: A Comprehensive Survey on Resources, Benchmarks, and Annotation Standards overlaps with clean result "Only continuous soft prefixes hit both EM axes at once on Qwen-2.5-7B-Instruct: discrete prompt searches split between the alignment objective and the distributional objective, and both discretizations of the soft prefix collapse (MODERATE confidence)", clean result "Training one persona to emit a [ZLT] marker without bystanders adopting it has a one-cell-wide LR x epochs window on Qwen2.5-7B-Instruct (LOW confidence)", clean result "The marker is a representational handle, not a behavioural one — sharing it between a villain persona and the assistant transfers no misalignment (HIGH confidence)". Matching terms: eval, source, line, trained, language. Source: arxiv cs.CL (NLP).
arXiv:2606.19352v1 Announce Type: new Abstract: Sign languages are expressive visual languages used by Deaf and Hard-of-Hearing (DHH) communities. Despite substantial progress in sign-language recognition, translation, and production, advances remain constrained by fragmented datasets, inconsistent annotations, and limited linguistic coverage. Existing benchmarks often fail to reflect real-world communication needs, and systematic analyses of these limitations remain limited. In this survey, we present a comprehensive index of sign-language datasets, covering 120 resources across 35 sign languages. We analyze key challenges such as modality imbalance, annotation granularity, and signer bias, and outline considerations for future dataset design. We also introduce a 24-field Sign-Language Datasheet and release a public GitHub repository (https://github.com/Ginqwerty/Open-Sign-Language) to support standardized documentation and reproducible evaluation. Overall, our work provides a unified and practical foundation for developing inclusive, robust, and scalable sign-language technologies in real-world applications.
Potential threat/caveat for clean result "Only continuous soft prefixes hit both EM axes at once on Qwen-2.5-7B-Instruct: discrete prompt searches split between the alignment objective and the distributional objective, and both discretizations of the soft prefix collapse (MODERATE confidence)": this item discusses limitation, limitations, bias, evaluation, benchmark.
- score 100arxiv cs.CL (NLP)arxiv:2606.19351unread
Detecting Hallucinations for Large Language Model-based Knowledge Graph Reasoning
Xinyan Zhu, Yaoqi Liu, Yue Gao, Huadong Ma, Cheng Yang, Chuan Shi · 2026-06-20
arXiv:2606. 19351v1 Announce Type: new Abstract: Knowledge graph (KG) reasoning infers new knowledge from existing facts and is widely applied in question answering, recommendation, and decision support.
Read next because Detecting Hallucinations for Large Language Model-based Knowledge Graph Reasoning overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: text, rect, correct, eval, line, rate, compare, contexts. Source: arxiv cs.CL (NLP).
arXiv:2606.19351v1 Announce Type: new Abstract: Knowledge graph (KG) reasoning infers new knowledge from existing facts and is widely applied in question answering, recommendation, and decision support. With the rapid development of large language models (LLMs), LLM-based KG reasoning frameworks have become increasingly popular by leveraging retrieved KG information. However, hallucinations in LLMs remain a critical issue. Even when relevant KG knowledge is incorporated, models may still generate incorrect outputs, leading to misinformation and unreliable decisions. Existing hallucination detection methods either focus on LLM internal states or verify consistency with retrieved contexts, but both overlook the structural information in KGs, resulting in suboptimal performance. To address this gap, we propose LUCID, the first halLUcination deteCtIon method for LLM-based knowleDge graph reasoning frameworks. LUCID jointly leverages LLM attention scores, KG semantics, and structural information. Specifically, it extracts node and edge features from attention scores and semantic similarities, and integrates them with KG structure using a graph neural network. We also construct manually annotated benchmark datasets for evaluation. Experiments on nine datasets show that LUCID achieves state of the art performance compared to 15 baselines.
Potential threat/caveat for clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)": this item discusses evaluation, benchmark.
- score 100arxiv cs.CL (NLP)arxiv:2606.19350unread
Pruning via Causal Attribution Preserves Reasoning Performance in Large Language Models
Amogh Sheth, Biruk Assefa, Yi Wen Huang, Andrew Lin, Yuhao Ge · 2026-06-20
arXiv:2606. 19350v1 Announce Type: new Abstract: Large language models (LLMs) excel at multi-step reasoning but incur substantial inference cost.
Read next because Pruning via Causal Attribution Preserves Reasoning Performance in Large Language Models overlaps with clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)", clean result "Coupling evil personas with wrong answers fails to protect Qwen2.5-7B from EM-induced alignment collapse — and the apparent capability ordering across coupling conditions is mostly eval contamination (LOW confidence)", clean result "Only continuous soft prefixes hit both EM axes at once on Qwen-2.5-7B-Instruct: discrete prompt searches split between the alignment objective and the distributional objective, and both discretizations of the soft prefix collapse (MODERATE confidence)". Matching terms: rect, arc-c, eval, rate, project, language, model. Source: arxiv cs.CL (NLP).
arXiv:2606.19350v1 Announce Type: new Abstract: Large language models (LLMs) excel at multi-step reasoning but incur substantial inference cost. We introduce Causal Attribution Pruning (CAP), a training-free method that identifies critical attention heads by measuring their causal impact on reasoning tasks and uses these head-level scores to guide fine-grained weight pruning. For each attention head, CAP estimates the expected performance degradation when the head is masked during forward passes on a small calibration set of reasoning problems. These causal scores are then converted into weight-level importance values for the corresponding projection matrices. Unlike magnitude-only or activation-based criteria, CAP's interventional measurement directly captures each head's functional contribution, yielding relative accuracy gains of up to 61% over Wanda on ARC-Challenge at 20% sparsity. We evaluate CAP on GSM8K, StrategyQA, and ARC-Challenge using Llama-3-8B-Instruct and Mistral-7B-Instruct at 10%, 20%, and 50% sparsity. At moderate sparsity (10-20%), CAP improves over Wanda in most model-benchmark configurations. with especially large gains on ARC-Challenge for Llama-3. Our results suggest that attention-head-level causal attribution can better preserve reasoning performance on downstream benchmarks than correlational pruning criteria at equivalent sparsity, while remaining limited by coarse MLP attribution at 50% sparsity.
Potential threat/caveat for clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)": this item discusses benchmark.
- score 100arxiv cs.CL (NLP)arxiv:2606.19349unread
Where to Place the Query? Unveiling and Mitigating Positional Bias in In-Context Learning for Diffusion LLMs via Decoding Dynamics
Zhengheng Li, Panrui Li, Xuyang Liu, Puzhi Xia · 2026-06-20
arXiv:2606. 19349v1 Announce Type: new Abstract: While In-Context Learning (ICL) is extensively studied in Autoregressive (AR) LLMs, its mechanism within Diffusion Large Language Models (dLLMs) remains largely unexplored.
Read next because Where to Place the Query? Unveiling and Mitigating Positional Bias in In-Context Learning for Diffusion LLMs via Decoding Dynamics overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: code, text, rect, line, rate, without, position, language. Source: arxiv cs.CL (NLP).
arXiv:2606.19349v1 Announce Type: new Abstract: While In-Context Learning (ICL) is extensively studied in Autoregressive (AR) LLMs, its mechanism within Diffusion Large Language Models (dLLMs) remains largely unexplored. Unlike AR models restricted by unidirectional causal masking, dLLMs intrinsically utilize bidirectional attention, offering extensive spatial flexibility for query placement. Unfortunately, current practices conventionally inherit AR-style trailing-query templates, often overlooking the structural paradigm shift. This paper presents a comprehensive analysis unveiling that query position is actually a first-order variable in dLLMs. Through empirical decoupling, we demonstrate that positional variance impacts generation quality on par with example semantic quality. Internally, this positional sensitivity stems from a spatial ``Recency Effect'' in attention flow and task-dependent shifts in decoding trajectories. To mitigate this instability without ground-truth labels, we reveal that traditional single-step confidence ($C_{decoded}$) fails in dLLMs. Instead, we propose Average Confidence ($\overline{C}$), a novel metric tracking the iterative decoding process. By establishing the foundational spatial ICL baselines, we introduce Auto-ICL, a training-free adaptive routing strategy that dynamically optimizes query placement, robustly approaching oracle performance across heterogeneous reasoning and perception tasks.
Potential threat/caveat for clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)": this item discusses bias.
- score 100arxiv cs.CL (NLP)arxiv:2606.19347unread
How LLMs Fail and Generalize in RTL Coding for Hardware Design?
Guan-Ting Liu, Chao-Han Huck Yang, Chenhui Deng, Zhongzhi Yu, Brucek Khailany, Yu-Chiang Frank Wang · 2026-06-20
arXiv:2606. 19347v1 Announce Type: new Abstract: Translating sequential programming priors into the parallel temporal logic of hardware design remains a crucial bottleneck for large language models(LLM).
Read next because How LLMs Fail and Generalize in RTL Coding for Hardware Design? overlaps with clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Coupling evil personas with wrong answers fails to protect Qwen2.5-7B from EM-induced alignment collapse — and the apparent capability ordering across coupling conditions is mostly eval contamination (LOW confidence)", clean result "Only continuous soft prefixes hit both EM axes at once on Qwen-2.5-7B-Instruct: discrete prompt searches split between the alignment objective and the distributional objective, and both discretizations of the soft prefix collapse (MODERATE confidence)". Matching terms: latin, alignment, eval, line, rate, test, language, model. Source: arxiv cs.CL (NLP).
arXiv:2606.19347v1 Announce Type: new Abstract: Translating sequential programming priors into the parallel temporal logic of hardware design remains a crucial bottleneck for large language models(LLM). To investigate this, we introduce a new error taxonomy grounded in problem solvability, inspired by cognitive theory. Our taxonomy categorizes failures into syntactic, semantic, solvable functional, and unsolvable functional types. Evaluations reveal a strict empirical ceiling on the VerilogEval benchmark, as frontier models plateau at a 90.8% initial pass rate. These plateaus are defined by unsolvable functional errors, exposing persistent knowledge gaps immune to test time compute scaling. Furthermore, we expose a striking surface convergence gap: optimization readily eliminates syntax errors but concurrently exacerbates deeper functional failures. Our findings demonstrate that alignment techniques merely teach models to compile. While repeated sampling strategies can patch solvable errors, register-transfer level(RTL) coding capacity remains strictly bounded by pretraining knowledge. Addressing challenges in the current LLM based hardware generation pipeline requires more studies in model reasoning rather than alignment interventions.
Potential threat/caveat for clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)": this item discusses failure, failures, evaluation, benchmark.
- score 100arxiv cs.CL (NLP)arxiv:2606.19344unread
Exposing the Unsaid: Visualizing Hidden LLM Bias through Stochastic Path Aggregation
Matteo Pelossi, Rita Sevastjanova, Thilo Spinner, Mennatallah El-Assady · 2026-06-20
arXiv:2606. 19344v1 Announce Type: new Abstract: Large Language Models (LLMs) exhibit representational and syntactic biases that are difficult to evaluate due to the stochastic nature of text generation.
Read next because Exposing the Unsaid: Visualizing Hidden LLM Bias through Stochastic Path Aggregation overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: text, class, rect, under, eval, token, line, full. Source: arxiv cs.CL (NLP).
arXiv:2606.19344v1 Announce Type: new Abstract: Large Language Models (LLMs) exhibit representational and syntactic biases that are difficult to evaluate due to the stochastic nature of text generation. Standard auditing methods rely on a single output inspection or static automated metrics. These approaches obscure the underlying probability distributions and fail to capture biases hidden in lower-probability generation branches. This paper introduces TreeTracer, a visual analytics tool designed to evaluate LLM bias through aggregated comparison. Using a systematic perturbation analysis pipeline, the tool replaces ontology-defined terms in each input prompt, aggregates hundreds of stochastic generations into a syntax-aligned hierarchical structure, and then performs classification-aware node merging with an auxiliary language model. The resulting structure is visualized through a custom Sankey diagram. By juxtaposing two ontology-driven trees, the workspace enables direct comparison between semantic contexts and supports systematic bias detection. Because any visualization reflects only a subset of the model's learned behavior, the system further applies contrastive inference to compute and directly display counterfactual token probabilities across contexts, reducing the risk of misinterpreting the presence of bias. We validate the workspace through case studies comparing an unaligned baseline model GPT-2 XL against the constitutionally aligned Apertus models. The visual aggregation successfully exposes hidden representational harms, such as counterfactual pronoun suppression and conversational marginalization of individuals. A preliminary user study confirms that the aggregated comparative interface reduces cognitive load and effectively supports analysts in detecting systemic biases.
Potential threat/caveat for clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)": this item discusses bias.
- score 64M7 QA inline RSS threat sourceunread
Artifact verification caveats for Sagan clean results
M7 QA · No release date
This paper studies failure modes and caveats when Sagan creates a clean result only after verifying an artifact row. It proposes benchmark checks for artifact verification, clean-result review comments, and negative controls.
Read next because Artifact verification caveats for Sagan clean results overlaps with experiment "Add C2 control arm (donor sees marker_B without marker_A) to disambiguate paired-marker binding from marker_B leaking alone". Matching terms: control. Source: M7 QA inline RSS threat source.
This paper studies failure modes and caveats when Sagan creates a clean result only after verifying an artifact row. It proposes benchmark checks for artifact verification, clean-result review comments, and negative controls.
Potential threat/caveat for experiment "Add C2 control arm (donor sees marker_B without marker_A) to disambiguate paired-marker binding from marker_B leaking alone": this item discusses failure, caveat, caveats, negative, benchmark.