Note A human preference dataset for multi-turn interleaved multimodal understanding and generatin tasks

Note Images associated with InterMT-Bench, a benchmark designed for evaluating MLLMs-as-a-Judge for multi-turn multimodal understanding and generation scenarios. Details refer to https://github.com/cby-pku/INTERMT

PKU-Alignment/InterMT-Judge

Updated 10 days ago • 7

Note A judge model designed to evaluate the quality of each turn in multi-turn dialogues, achieving high agreement rate with human assessments.