laion/moss-va-v2-sft2-eval
1
Two-model comparison: SFT2 baseline vs SFT2+DPO2, same 80 prompts
Add SFT2+DPO2 players side by side with the SFT2 baseline
Round-2 SFT evaluation: 40 emotions x EN/DE, timed-script prompts, players
initial commit
Two-model comparison: SFT2 baseline vs SFT2+DPO2, same 80 prompts
Add SFT2+DPO2 players side by side with the SFT2 baseline
Round-2 SFT evaluation: 40 emotions x EN/DE, timed-script prompts, players
initial commit