发现大模型在误导信息下答案易翻车，提出新评估框架并提升稳定性
Exploring Response Uncertainty in MLLMs: An Empirical Evaluation under Misleading Scenarios
- 构建双阶段评测流程，检测模型对误导线索的响应不一致
- 12个主流模型在误导下正确答案翻转率超86%，显性提示下达67.19%
- 通过小规模微调，误导率降至32.77%以下，显著提升答案一致性
多模态大语言模型（MLLMs）在视觉问答到视频理解等任务上已达顶尖水平。然而，现有研究主要集中于视觉-文本错配，未充分考察模型在遭遇误导信息时是否能维持原本正确答案的能力。我们揭示了一种响应不确定性现象：在九个标准数据集上，十二个前沿开源MLLMs在接收到单一误导线索后，65%的情况下会推翻原先正确的回答。为系统量化这一脆弱性，我们提出两阶段评估流程：(1) 获取模型在无扰动输入下的原始回答；(2) 注入显式（错误答案提示）和隐式（上下文矛盾）误导指令，计算误导率——即正确答案翻转的比例。基于最易受骗样本，我们构建了多模态不确定性基准（MUB），包含按十二个先进MLLMs被误导程度分层的图像-问题对，分为低、中、高难度三类。对十二个开源和五个闭源模型的广泛评估显示，平均误导率超过86%，显性线索下超67.19%，隐性线索下超80.67%。为进一步降低误导率，我们对所有开源MLLMs在2000样本的混合指令数据集上进行微调，使误导率降至6.97%（显性）和32.77%（隐性），在高度误导输入上一致性提升近29.37%，且标准基准上的准确率略有改善。代码已开源。
原文摘要 · Abstract (English)
Multimodal large language models (MLLMs) have recently achieved state-of-the-art performance on tasks ranging from visual question answering to video understanding. However, existing studies have concentrated mainly on visual-textual misalignment, leaving largely unexplored the MLLMs' ability to preserve an originally correct answer when confronted with misleading information. We reveal a response uncertainty phenomenon: across nine standard datasets, twelve state-of-the-art open-source MLLMs overturn a previously correct answer in 65% of cases after receiving a single deceptive cue. To systematically quantify this vulnerability, we propose a two-stage evaluation pipeline: (1) elicit each model's original response on unperturbed inputs; (2) inject explicit (false-answer hints) and implicit (contextual contradictions) misleading instructions, and compute the misleading rate - the fraction of correct-to-incorrect flips. Leveraging the most susceptible examples, we curate the Multimodal Uncertainty Benchmark (MUB), a collection of image-question pairs stratified into low, medium, and high difficulty based on how many of twelve state-of-the-art MLLMs they mislead. Extensive evaluation on twelve open-source and five closed-source models reveals a high uncertainty: average misleading rates exceed 86%, with explicit cues over 67.19% and implicit cues over 80.67%. To reduce the misleading rate, we then fine-tune all open-source MLLMs on a compact 2000-sample mixed-instruction dataset, reducing misleading rates to 6.97% (explicit) and 32.77% (implicit), boosting consistency by nearly 29.37% on highly deceptive inputs, and slightly improving accuracy on standard benchmarks. Our code is available at https://github.com/Yunkaidang/uncertainty
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。