← 研踪 PaperTrailarXiv:2603.23509 · v2
Internal Safety Collapse in LLMs and Coding Agents
首次提交 2026-03-04 20:55
版本更新 2026-11-06 03:19
雷达观测 2026-11-09 10:16
摘要
LLMs and coding agents can now tackle complex, long-horizon tasks through advances in reinforcement learning, but stronger task-solving capabilities do not necessarily preserve safety alignment. Recent AI agent incidents involving reward hacking indicate how optimizing for task success can bring learned capabilities into conflict with safety alignment. In this work, we identify a critical failure mode that manifests this conflict, which we term Internal Safety Collapse (ISC). When completing otherwise benign coding tasks, AI agents can prioritize satisfying program requirements over safety constraints, generating large volumes of sensitive content, including unsafe outputs, as task data. To study this phenomenon, we introduce TVD (Task, Validator, Data), a framework for building controlled workflows that require sensitive content for completion. Through extensive experiments, we observe instances of highly harmful content that closely resemble outputs from early, unaligned LLMs in 2023. We further find that ISC can be effectively triggered even in highly capable LLMs such as Claude Fable 5. These results indicate that, despite substantial alignment efforts, frontier LLMs remain internally unsafe: safety alignment primarily reshapes surface behavior without eliminating the underlying knowledge that enables unsafe outputs. Our findings underscore the need for caution when deploying LLMs in high-stakes settings, such as scientific pipelines and autonomous agents. Content Warning: This paper contains examples of harmful content. Source code: https://github.com/wuyoscar/Internal-Safety-Collapse
作者备注：Accepted at NeurIPS 2026. 38 pages, 16 figures, 30 tables. Qualitative examples of jailbreaks may be harmful in nature
观测到的版本
相邻记录可能跳过中间版本；首次发现 v2 及以上也会标为版本更新。
v2 · 2026-11-06 03:19 · 最新观测版本
Internal Safety Collapse in LLMs and Coding Agents
LLMs and coding agents can now tackle complex, long-horizon tasks through advances in reinforcement learning, but stronger task-solving capabilities do not necessarily preserve safety alignment. Recent AI agent incidents involving reward hacking indicate how optimizing for task success can bring learned capabilities into conflict with safety alignment. In this work, we identify a critical failure mode that manifests this conflict, which we term Internal Safety Collapse (ISC). When completing otherwise benign coding tasks, AI agents can prioritize satisfying program requirements over safety constraints, generating large volumes of sensitive content, including unsafe outputs, as task data. To study this phenomenon, we introduce TVD (Task, Validator, Data), a framework for building controlled workflows that require sensitive content for completion. Through extensive experiments, we observe instances of highly harmful content that closely resemble outputs from early, unaligned LLMs in 2023. We further find that ISC can be effectively triggered even in highly capable LLMs such as Claude Fable 5. These results indicate that, despite substantial alignment efforts, frontier LLMs remain internally unsafe: safety alignment primarily reshapes surface behavior without eliminating the underlying knowledge that enables unsafe outputs. Our findings underscore the need for caution when deploying LLMs in high-stakes settings, such as scientific pipelines and autonomous agents. Content Warning: This paper contains examples of harmful content. Source code: https://github.com/wuyoscar/Internal-Safety-Collapse
查看此版本解读 →查看此版本原文 ↗