finding
active
finding:average-kl-between-base-and-policy-is-10-20-larger-on-not-supported-than-supported-tokens

Average KL between base and policy is 10–20% larger on Not Supported than Supported tokens

Targeted updates.

Source paper

extracted_from
Features as Rewards: Scalable Supervision for Open-Ended Tasks via Interpretability
(2026) · Aaditya Vikram Prasad · Connor Watts · Jack Merullo · Dhruvil Gala +3

Neighborhood — ranked by edge-count