Paper

LoMC: Localized Multidirectional Correction for Refusal Suppression in Routed Foundation Models

Unreadunread

AI summary

arXiv:2606. 13709v1 Announce Type: new Abstract: We study controlled post-training refusal suppression in routed MoE and hybrid-MoE foundation models, aiming to increase non-refusal target-response behavior while preserving general capability under a compact intervention footprint.

Reading status

Priority

About this paper

Authors: Yan Hong, Kedong Xiu, Wei Li, Jun Lan, Huijia Zhu, Shuheng Zhou, Zhongcai Lyu, Weiqiang Wang, Jianfu Zhang
Released: 2026-06-15 (24d ago)
Topic: Related to current project
Relation to my work: Read next because LoMC: Localized Multidirectional Correction for Refusal Suppression in Routed Foundation 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: text, rect, under, correct, control, without, capability, model. Source: arxiv stat.ML (Machine Learning).
arXiv: 2606.13709

Ask Claude about this paper

Posts a comment with @claude so the runner spins up a Claude Code session scoped to LoMC: Localized Multidirectional Correction for Refusal Suppression in Routed Foundation Models. Replies appear in the conversation below.

Comments

No comments yet.

Related records

No links yet.