Skip to main content
Search

# Chain-of-Thought Monitorability

1 related articles

Subscribe to updates
  1. arXiv

    RLVR post-training is proved to permit unbounded language drift while supervised fine-tuning does not, with drift appearing only on novel reasoning tasks the base model cannot perform

    This work characterizes when language drift arises in RLVR post-training: it proves theoretically that RLVR optimization pressure permits unbounded language drift whereas supervised fine-tuning does not, shows empirically that drift arises specifically during RLVR on novel reasoning tasks where the target behavior cannot be drawn out of the base model, and further proves that language drift cannot be constrained without constraining expected reward.

You’re all caught up