List your company

Dialogue on Dialogue (test)

8.84FluencySFT
8.3728.49358.6158.7365Jun 21, 2024

Evaluation results

MethodLinks
SFTBackbone=LLaMA2-13B2024.068.847.778.497.437.31
HBATBackbone=LLaMA2-13B, Optimization=PPO2024.068.87.78.487.397.89
HBATBackbone=LLaMA2-13B, Optimization=DPO2024.068.797.88.457.517.96
DPOBackbone=LLaMA2-13B, Optimization=DPO2024.068.677.138.547.637.84
RLHFBackbone=LLaMA2-13B, Optimization=PPO2024.068.397.628.877.477.72