感知还是偏见：大语言模型能否超越个性的第一印象？
Perception or Prejudice: Can MLLMs Go Beyond First Impressions of Personality?
AI总结：
本文探讨了多模态大语言模型（MLLMs）在感知个性方面的能力，提出了一种新的任务Grounded Personality Reasoning（GPR），并构建了一个新的数据集MM-OCEAN，通过三重评估体系揭示了MLLMs在人格推理中的偏见问题。
摘要：
多模态大语言模型（MLLMs）正在越来越多地应用于需要感知个性的人类交互角色中，但现有的基准测试仅评估其对大五人格特质分数的预测能力，未能确定模型是通过行为理解真正感知个性，还是仅通过表面模式匹配进行偏见判断。我们通过三个贡献填补了这一空白：（i）一个新的任务：我们正式定义了Grounded Personality Reasoning（GPR），要求MLLMs通过一系列评分、推理和锚定过程，将每个大五评分与可观察的证据联系起来；（ii）一个新的数据集：我们发布了MM-OCEAN（1,104个视频，5,320个多项选择题），由多代理流程生成，包含时间戳行为观察、证据支持的特质分析以及七类线索锚定多项选择题；（iii）基准测试和分析：我们设计了一个三级评估体系（评分、推理、锚定）以及四个样本级失败模式指标：偏见率（PR）、编造率（CR）、整合失败率（IR）和整体锚定率（HR），并基准测试了27个MLLMs（13个封闭式，14个开放式）。分析揭示了一个显著的偏见差距：在所有正确评分中，51%的评分没有基于检索到的线索进行锚定，而整体锚定率仅在0-33.5%之间。这些发现揭示了获得正确分数与为正确原因推理之间的脱节，为MLLMs中的扎根社会认知绘制了路线图。
英文摘要：
Multimodal Large Language Models (MLLMs) are increasingly deployed in human-facing roles where personality perception is critical, yet existing benchmarks evaluate this capability solely on numerical Big Five score prediction, leaving open whether models truly perceive personality through behavioral understanding or merely prejudge through superficial pattern matching. We address this gap with three contributions. (i) A new task: we formalize Grounded Personality Reasoning (GPR), which requires MLLMs to anchor each Big Five rating in observable evidence through a chain of rating, reasoning, and grounding. (ii) A new dataset: we release MM-OCEAN (1,104 videos, 5,320 MCQs), produced by a multi-agent pipeline with human verification, with timestamped behavioral observations, evidence-grounded trait analyses, and seven categories of cue-grounding MCQs. (iii) Benchmark and analysis: we design a three-tier evaluation (rating, reasoning, grounding) plus four sample-level failure-mode metrics: Prejudice Rate (PR), Confabulation Rate (CR), Integration-failure Rate (IR), and Holistic-grounding Rate (HR), and benchmark 27 MLLMs (13 closed, 14 open). The analysis uncovers a striking Prejudice Gap: across the field, 51% of correct ratings are not grounded in retrieved cues, and the Holistic-Grounding Rate spans only 0-33.5%. These findings expose a disconnect between getting the right score and reasoning for the right reason, charting a roadmap for grounded social cognition in MLLMs.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
