MT-RL-Judge uses multi-task RL to train MLLM judges that outperform baselines in consistency, human preference correlation, and out-of-distribution generalization.
G-eval: Nlg evaluation using gpt-4 with better human alignment
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Multi-Task Reinforcement Learning for Enhanced Multimodal LLM-as-a-Judge
MT-RL-Judge uses multi-task RL to train MLLM judges that outperform baselines in consistency, human preference correlation, and out-of-distribution generalization.