LLaVA-Critic：学习评估多模型

LLaVA-Critic: Learning to Evaluate Multimodal Models

October 3, 2024

作者: Tianyi Xiong, Xiyao Wang, Dong Guo, Qinghao Ye, Haoqi Fan, Quanquan Gu, Heng Huang, Chunyuan Li

cs.AI

摘要

我们介绍了LLaVA-Critic，这是第一个开源的大型多模态模型（LMM），旨在作为通用评估器，评估广泛的多模态任务性能。LLaVA-Critic使用高质量的评论指令跟随数据集进行训练，该数据集包含多样的评估标准和场景。我们的实验表明该模型在两个关键领域的有效性：（1）LMM作为评判者，在多个评估基准上，LLaVA-Critic提供可靠的评估分数，表现与或超过GPT模型；（2）偏好学习，它为偏好学习生成奖励信号，增强模型对齐能力。这项工作强调了开源LMM在自我批评和评估中的潜力，为未来研究提供了舞台，探讨LMM的可扩展、超人类对齐反馈机制。

English

We introduce LLaVA-Critic, the first open-source large multimodal model (LMM) designed as a generalist evaluator to assess performance across a wide range of multimodal tasks. LLaVA-Critic is trained using a high-quality critic instruction-following dataset that incorporates diverse evaluation criteria and scenarios. Our experiments demonstrate the model's effectiveness in two key areas: (1) LMM-as-a-Judge, where LLaVA-Critic provides reliable evaluation scores, performing on par with or surpassing GPT models on multiple evaluation benchmarks; and (2) Preference Learning, where it generates reward signals for preference learning, enhancing model alignment capabilities. This work underscores the potential of open-source LMMs in self-critique and evaluation, setting the stage for future research into scalable, superhuman alignment feedback mechanisms for LMMs.