Guiding LLM Post-training Data Engineering with Model Internals from Sparse Autoencoders 事件

Guiding LLM Post-training Data Engineering with Model Internals from Sparse Autoencoders arXiv:2605.27354v1 Announce Type: cross Abstract: Model internals encode rich information about how a large language model (LLM) processes its training data; however, post-training data engineering largely relies on external signals and ignores rich intrinsic signals lying in model internals. We propose SAERL, a data engineering framework for LLM reinforcement learning (RL). It models three intrinsic data p