List your company

Cumulative temporal regret on CAP-GSM8K

21.3RLVR R_temp CERLVR
12.273.625135.05196.475Jun 3, 2026

Evaluation results

MethodLinks
RLVRModel=Claude-Sonnet-4.5, E=100, seeds=32026.0621.3——
RLVRModel=GPT-4o, E=100, seeds=32026.0658——
RLVRModel=GPT-3.5-Turbo, E=100, seeds=32026.0668——
RLVRModel=Llama-3.3-70b, E=500, seeds=102026.06248.8——
TriviumModel=Llama-3.3-70b, E=500, seeds=102026.06—2.4103
TriviumModel=GPT-4o, E=100, seeds=32026.06—319
TriviumModel=Claude-Sonnet-4.5, E=100, seeds=32026.06—2.39
TriviumModel=GPT-3.5-Turbo, E=100, seeds=32026.06—3.718