Compute as Teacher: Turning Inference Compute Into Reference-Free Supervision — mentioned by NitpickLawyer

Thu, 18 Sep 2025 15:55:35 GMT

Paper link: https://arxiv.org/abs/2509.14234 Some interesting tidbits. - they propose several "judges", each with their own model (weights at different stages) and separate "concerns". The generate part evolves with the model (in RL) while the "gather and reconcile" is fixed at a frozen stage. - the "gather and reconcile" judge doesn't get the question when analysing the entire rollout set! (I hope I read this correctly "We keep the anchor question-blind to prevent it from acting as just another rollout and to encourage genuine cross-rollout reasoning") - a 2nd judge "marks" binary yes/no…

Paper and discussion history. A mention is not an endorsement.

Compute as Teacher: Turning Inference Compute Into Reference-Free Supervision · HackerPapers discussions

Compute as Teacher: Turning Inference Compute Into Reference-Free Supervision — mentioned by NitpickLawyer