Paper proposes stage-specific data sets for SFT-then-RL in SLMs

Arxiv CS.CL (NLP/language models)

The paper argues that data strategy should align with the distinct roles of SFT and RL stages. It proposes stage-specific data sets for small language model reasoning post-training.

llmreasoning

Arxiv CS.CL (NLP/language models)·Chongyang HeRui ZhangZixuan WangXin Li

7 days ago