List your company

Preference Evaluation on PKU-SafeRLHF

57Win RateDLMA-13B
41.445.4549.553.55Feb 19, 2024

Evaluation results

MethodLinks
DLMA-13BComparison Baseline=Llama2-13B2024.0257835
DLMA-7BComparison Baseline=RLAIF-7B2024.0256836
DLMA-7BComparison Baseline=Llama2-7B2024.0255837
DLMA-13BComparison Baseline=RLAIF-13B2024.02551134
DLMA-13BComparison Baseline=CD-13B2024.02491645
DLMA-7BComparison Baseline=RLCD-7B2024.02432532
DLMA-13BComparison Baseline=RLCD-13B2024.02432433
DLMA-7BComparison Baseline=CD-7B2024.02421543