List your company

Helpfulness on HH-RLHF (test)

2.542RewardMAVIS
2.086482.204742.3232.44126Aug 19, 2025

Evaluation results

MethodLinks
MAVISModel Size=13B2025.082.54224.11
PPOModel Size=13B2025.082.45224.19
MAVISModel Size=7B2025.082.31118.64
PPOModel Size=7B2025.082.10417.81