List your company

Binary Classification on ImageReward (test)

65.07Macro-F1RL-Single
54.6757.3760.0762.77Mar 12, 2026

Evaluation results

MethodLinks
RL-Singletraining=Single-task RL2026.0365.07
MT-RL-Judgetraining=Multi-task RL2026.0364.97
SFT-Singletraining=Single-task SFT2026.0364.95
SFT-Unifiedtraining=Unified multi-task SFT2026.0363.34
Off-the-shelf2026.0355.07