Paper

HARP: Measuring Harm Amplification in Multi-Agent LLM Systems

Unreadunread

AI summary

arXiv:2605. 27489v1 Announce Type: new Abstract: Multi-agent LLM systems decompose workflows across agents, tools, shared context, memory, and decision gates.

Reading status

Priority

About this paper

Authors: Md Hafizur Rahman, Zafaryab Haider, Tanzim Mahfuz, Prabuddha Chakraborty
Released: 2026-05-28 (1mo ago)
Topic: Related to current project
Relation to my work: Read next because HARP: Measuring Harm Amplification in Multi-Agent LLM 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, text, eval, token, rate, compare, full. Source: arxiv cs.CR (Cryptography and Security).
arXiv: 2605.27489

Ask Claude about this paper

Posts a comment with @claude so the runner spins up a Claude Code session scoped to HARP: Measuring Harm Amplification in Multi-Agent LLM Systems. Replies appear in the conversation below.

Comments

No comments yet.

Related records

No links yet.