arXiv NLP
techCenter
Reward-Informed Sparse Autoencoders and the Solution-Completeness Confoundtranslating…
1 min readUnknownarXiv Press Group
arXiv:2608.26136v1 Announce Type: new
Abstract: Sparse autoencoders (SAEs) decompose language-model activations into sparse, interpretable features, and an appealing way to aim them at reasoning is to curate their data with a signal reinforcement learning already produces: the…