List your company

LLM Alignment on Gemma-3-4B

94.33Win RateNash Prox
2.123626.06185073.9382May 26, 2025

Evaluation results

MethodLinks
Nash ProxOpponent=SFT2025.0594.33
Online IPOOpponent=SFT2025.0593.64
Online DPOOpponent=SFT2025.0593.26
Nash ProxOpponent=Online IPO2025.0556.27
Nash ProxOpponent=Online DPO2025.0553.44
Online DPOOpponent=Online IPO2025.0552.83
Online IPOOpponent=Online DPO2025.0547.17
Online DPOOpponent=Nash Prox2025.0546.56
Online IPOOpponent=Nash Prox2025.0543.73
SFTOpponent=Online DPO2025.056.74
SFTOpponent=Online IPO2025.056.36
SFTOpponent=Nash Prox2025.055.67