Reward modeling
Technique · Training · 1 item since Apr 2026
Training models that score outputs, supplying the reward signal for RL.
Trends
Items
- Apr 15
C2 learns rubric reward models from preferences
Jointly training a cooperative rubric generator and a critical verifier from binary preferences alone lets an 8B reward model match a 32B rubric-annotated baseline, gaining 6.5 points on RM-Bench and 6.0 points of length-controlled win rate on AlpacaEval 2.0.