List your company

CVaR(0.125) Combined Win Rate Evaluation on PKU-SafeRLHF (n=100 Conflict Prompts)

40.3CVaR(0.125) Combined Win RateIPO Entropic τ=10
0.88411.11721.3531.583May 11, 2026

Evaluation results

MethodLinks
IPO Entropic τ=10tau (τ)=10, objective=Entropic2026.0540.3
IPO CVaR α=0.25alpha (α)=0.25, objective=CVaR2026.0537.2
IPO Entropic τ=5tau (τ)=5, objective=Entropic2026.0537.1
IPO CVaR α=0.125alpha (α)=0.125, objective=CVaR2026.0534.2
EG Entropic τ=5tau (τ)=5, objective=Entropic2026.0534.1
IPO k = 1 Neutralk=1, objective=Neutral2026.0533.8
EG k = 1 Neutralk=1, objective=Neutral2026.0532.4
NMD k = 1 Neutralk=1, objective=Neutral2026.0528
EGPO† (HF)source=Hugging Face2026.054.8
SFT Basetype=Baseline2026.052.4