# 重新审视后训练中的完整推理轨迹：部分轨迹与端点训练更有效

- 来源：HuggingFace Daily Papers（社区热门论文）
- 发布时间：2026-09-07 08:00
- AIHOT 分数：49
- AIHOT 链接：https://aihot.news/items/cmtuvps820gamrorpegy1nysl
- 原文链接：https://arxiv.org/abs/2609.07103

## AI 摘要

研究发现，LLM 后训练（如 SFT）中完整推理轨迹带来的收益有限，而部分轨迹即使被大幅截断仍然有效。基于注意力的分析和受控 token 移除实验表明，中间 token 对最终推理质量贡献极小，模型可依据已知轨迹端点从内部知识推断缺失步骤。仅用端点训练还能改变推理行为，并提升基于强化学习或 on-policy 蒸馏的后训练方法效果。

## 正文

Large language models (LLMs) are often post-trained on pre-collected reasoning trajectories to improve their reasoning capability. Such trajectories tend to be long due to complex, interwoven paths, which often include detours on the path toward the answer. However, it has been underexplored whether LLMs indeed benefit from learning complete trajectories in post-training, such as supervised fine-tuning (SFT). Starting from our pilot study, we find that full trajectories provide only limited benefit, while partial trajectories are effective even under heavy truncation. We analyze redundancy in reasoning trajectories through attention-based analyses and controlled token-removal studies, both of which show that intermediate tokens contribute minimally to final reasoning quality. This suggests that avoiding redundant information may allow LLMs to internally infer coherent alternatives by inferring missing steps from their internal knowledge, given known trajectory endpoints. Furthermore, we show that training LLMs using endpoints leads to consistent changes in reasoning behavior, and that it also benefits post-training methods based on reinforcement learning or on-policy distillation, highlighting the need to revisit complete reasoning traces. Code is available at https://github.com/naver-ai/revisiting-trace.