List your company

Harmlessness on HH-RLHF (test)

2.772RewardMAVIS
2.446482.530992.61552.70001Aug 19, 2025

Evaluation results

MethodLinks
MAVISModel Size=13B2025.082.77211.32
PPOModel Size=13B2025.082.76211.4
MAVISModel Size=7B2025.082.6564.65
PPOModel Size=7B2025.082.4594.23