跳到正文
arXiv · cs.AI · Yiheng Wang, Yueqian Lin, Lichen Zhu, Yudong Liu, Hai "Helen" Li, Yiran Chen · 9 小时前模型评分44

当没有正确答案时:诊断多模态大模型在视频理解中的缺失答案检测

When No Answer Is Correct: Diagnosing Absent Answer Detection for MLLMs in Video Understanding

导读

一项诊断研究发现,多模态大模型(多模态大语言模型(MLLMs,Multimodal Large Language Models),Multimodal Large Language Models)在视频理解中面对候选选项均不正确的情况时,绝大多数会选出看似合理的干扰项,而非识别出没有有效答案。该失败在时间推理任务中更明显,且随帧采样变密而加重;思维链提示虽能显著提升检测率,但整体表现仍不理想。

生成六维分析
分类置信度

中 · 待人工复核

  • 只有单一媒体来源,需要人工复核。

来源级别:媒体与个人

事件类型:研究发布

证据摘录

  • “This work presents a diagnostic study of absent answer detection for MLLMs in video understanding, where the correct answer is deliberately excluded from the candidate set and a reliable model is expected to recognize that no valid option exists.”

来源:arXiv · cs.AI · arxiv.org