List your company

Reward Performance on Twenty Questions Low-data

3.46Test-Time RewardMC
0.25681.08841.922.7516May 23, 2026

Evaluation results

MethodLinks
MCtraining_regime=BOOST2026.053.46
MCtraining_regime=Bilevel Only2026.052.9
MCtraining_regime=Vanilla2026.051.01
MCtraining_regime=Vanilla+Syn2026.051.01
ILQLtraining_regime=BOOST2026.050.53
ILQLtraining_regime=Bilevel Only2026.050.51
ILQLtraining_regime=Vanilla+Syn2026.050.45
ILQLtraining_regime=Vanilla2026.050.38