List your company

Binary Classification on AGIN-Nat. (test)

81.75Macro-F1SFT-Unified
67.439671.154874.8778.5852Mar 12, 2026

Evaluation results

MethodLinks
SFT-Unifiedtraining=Unified multi-task SFT2026.0381.75
MT-RL-Judgetraining=Multi-task RL2026.0381.63
RL-Singletraining=Single-task RL2026.0380.5
SFT-Singletraining=Single-task SFT2026.0378.64
Off-the-shelf2026.0367.99