@inproceedings{zhou-etal-2026-cctvbench,
    title = "{CCTVB}ench: Contrastive Consistency Traffic {V}ideo{QA} Benchmark for Multimodal {LLM}s",
    author = "Zhou, Xingcheng  and
      Guo, Hao  and
      Song, Rui  and
      Zimmer, Walter  and
      Liu, Mingyu  and
      Schamschurko, Andr{\textbackslash}'e  and
      Cao, Hu  and
      Knoll, Alois",
    editor = "Liakata, Maria  and
      Moreira, Viviane P.  and
      Zhang, Jiajun  and
      Jurgens, David",
    booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {ACL} 2026",
    month = jul,
    year = "2026",
    address = "San Diego, California, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-acl/2026.findings-acl.1089/",
    pages = "21665--21684",
    ISBN = "979-8-89176-395-1",
    abstract = "Safety-critical traffic reasoning requires contrastive consistency: models must detect true hazards when an accident occurs, and reliably reject plausible-but-false hypotheses under near-identical counterfactual scenes. We present CCTVBench, a Contrastive Consistency Traffic VideoQA Benchmark built on paired real accident videos and world-model-generated counterfactual counterparts, together with minimally different, mutually exclusive hypothesis questions. CCTVBench enforces a single structured decision pattern over each video question quadruple and provides actionable diagnostics that decompose failures into positive omission, positive swap, negative hallucination, and mutual-exclusivity violation, while separating video versus question consistency. Experiments across open-source and proprietary video LLMs reveal a large and persistent gap between standard per-instance QA metrics and quadruple-level contrastive consistency, with unreliable none-of-the-above rejection as a key bottleneck. Finally, we introduce C-TCD, which leverages the semantically exclusive counterpart video as the contrast input at inference time, improving both instance-level QA and contrastive consistency."
}