Revisiting Complete Reasoning Traces for Post-Training
Researchers show full reasoning traces provide limited benefit in LLM post-training, with heavily truncated or endpoint-only trajectories performing comparably.
A pilot study plus attention-based analyses and controlled token-removal studies show intermediate tokens in reasoning trajectories contribute minimally to final reasoning quality. Partial trajectories remain effective even under heavy truncation, and training on endpoints alone leads to consistent changes in reasoning behavior. The finding also benefits reinforcement-learning and on-policy distillation post-training; code is released at github.com/naver-ai/revisiting-trace.
- Full trajectories give limited SFT benefit; partial trajectories remain effective even under heavy truncation.
- Attention analyses and token-removal studies show intermediate tokens minimally affect final reasoning quality.
- Endpoint-only training consistently changes reasoning behavior, implying models infer missing steps internally.
- Benefits extend to RL and on-policy distillation post-training; code released on GitHub.
From the original article
1 min read at huggingface.co
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).
By Jaehui Hwang, Sangdoo Yun, Byeongho Heo · huggingface.co
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.