Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Helpfulness Evaluation on LLaVA-Bench

93.1Conversation Score

LLaVA-RLHF

75.10879.77984.4589.121Jan 16, 2025
Updated 1mo ago

Evaluation Results

MethodLinks
2025.01
93.176.2105.691.8
2025.01
93.175.391.686.7
2025.01
87.178.390.785.5
2025.01
84.977.390.384.3
2025.01
84.174.489.883
2025.01
84.175.3106.888.9
2025.01
82.179.587.983.2
2025.01
80.774.588.481.4
2025.01
79.677.391.482.9
2025.01
7671.888.280.5
2025.01
75.883.790.783.5