List your company

Multi-turn Dialogue Alignment on ALOE (test)

2.4Rank-1 ScoreSFT
1.9842.0922.22.308Oct 13, 2025

Evaluation results

MethodLinks
SFT2025.102.433.83.83.8444.23.64.23.68
TPO2025.102.43.44.23.84.24.24.244.243.86
CoT2025.1023.23.844.24.44.23.83.83.83.72
GRPO2025.1023.433.43.63.23.43.433.43.18
CDRA2025.1023.444.24.44.44.64.2443.92