Tag
该论文对基于 LLM 的智能体系统的自动失败归因（AFA）进行了机制研究，发现模型存在'知道但不说'的差距：内部激活包含正确归因的信息，但无法显式输出。作者据此提出 grafting 干预方法，在多个 LLM 和基准上优于直接提示方法。
This paper introduces the Linear Accessibility Profile (LAP), a diagnostic method using logit lens to predict steering vector effectiveness across model layers, achieving ρ=+0.86 to +0.91 correlation on 24 concept families across five models. The work provides a systematic framework to determine which layers and concepts are suitable for steering interventions, replacing ad-hoc trial-and-error approaches.