arXiv cs.AI· Min-Seon Kim, Seungdong Yoa, Sanghyu Yoon, Ye Seul Sim, Suhee Yoon, Dongmin Kim, Jung-Eun Kim·· 2 小时前AI 评分50
arXiv 论文：LLM agent 蒸馏中思考监督并非总是必要
Agents, Think Only When it Helps
AI 导读
arXiv 论文 2611.05709 研究了 LLM agent 蒸馏中思考监督的作用，通过在蒸馏时独立扰动教师轨迹中的思考与动作来解耦两者影响。在多个常用 agent 基准和不同模型规模上，改变或丢弃思考对学生性能影响很小；只有当训练中动作未披露相应思考的信息时，思考监督才变得重要，作者据此提出可按信息是否披露来选择是否蒸馏思考。
正文 · 原文
Abstract:LLM agent distillation transfers agentic capabilities to student models by training on complete teacher trajectories that typically interleave thoughts and actions. We question how critical thought supervision is for agents' ultimate action performance. To answer this, we disentangle their impacts by independently perturbing teacher thoughts and actions during distillation. On widely adopted agent benchmarks across model scales, altering or discarding thoughts has little effect on student performance. We find, however, that thought supervision becomes important when information about the thought is not disclosed in the corresponding action during training. By identifying such discrepant cases, our findings challenge the common practice of distilling entire teacher trajectories for LLM agents, and offer a new insight that thought supervision can be a discernible choice for efficiency depending on whether the information of thought is disclosed in the action.
| Subjects: | Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | arXiv:2611.05709 [cs.AI] |
| (or arXiv:2611.05709v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2611.05709 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Jung-Eun Kim [view email]
[v1]
Tue, 3 Nov 2026 22:35:21 UTC (70 KB)
来源：arXiv cs.AI · arxiv.org