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

Multimodal Conversation on MMRole OOD

91.6LLM-as-a-Judge Score

Dr.GRPO (Latent Action)

67.78473.96780.1586.333Jan 12, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.01
91.6
2026.01
91.5
2026.01
90.1
2026.01
88.9
2026.01
87.2
2026.01
86.4
2026.01
86.3
2026.01
85.9
2026.01
85.6
2026.01
85.4
2026.01
84.2
2026.01
84.2
2026.01
84
2026.01
82.3
2026.01
82.1
2026.01
80.9
2026.01
80.5
2026.01
80.1
2026.01
79.6
2026.01
68.7