Metraipolis › archive › dc_e3218bbc5e2d

Can We Trust Video Hallucination Detectors? VidHalLoc for Evaluating the Evaluators — arXiv:2609.09895

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.09895
License (as archived)
CC0-1.0
Archived
2026-10-16T10:59:52.519Z
SHA-256 of the archived text
2cb912254fb38a6249fadfb8f65b432a1733e3b235c07a1fd7292258a0ffceb5
Stable URL of this page
https://metraipolis.com/documento/dc_e3218bbc5e2d

How to cite

Can We Trust Video Hallucination Detectors? VidHalLoc for Evaluating the Evaluators — arXiv:2609.09895 — arXiv (en), https://arxiv.org/abs/2609.09895 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_e3218bbc5e2d (Metraipolis archive dc_e3218bbc5e2d, sha256 2cb912254fb38a6249fadfb8f65b432a1733e3b235c07a1fd7292258a0ffceb5)
BibTeX
@misc{dce3218bbc5e2d,
  title = {Can We Trust Video Hallucination Detectors? VidHalLoc for Evaluating the Evaluators — arXiv:2609.09895},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.09895},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_e3218bbc5e2d; sha256 2cb912254fb38a6249fadfb8f65b432a1733e3b235c07a1fd7292258a0ffceb5},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.09895. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Video-language models and video agents can produce hallucinations that conflict with spatiotemporal evidence. Existing benchmarks mainly evaluate model hallucinations, and heterogeneous mechanisms make detector reliability difficult to compare. We introduce VidHalLoc, a benchmark that evaluates hallucination detection methods under a unified diagnostic evaluation protocol using 2,000 adversarial hallucination samples across Video Question Answering and Video Captioning tasks, spanning Ontology and Dynamic hallucination categories. To construct VidHalLoc efficiently, we introduce VideoHALO, a Harness Engineering-informed multi-agent workflow that decomposes data construction into four executable stages supported by a memory system and a communication protocol. Evaluation of fifteen methods reveals that the four dedicated detectors peak at an Overall accuracy of only 34.63%, indicating limited reliability across video hallucination types [Dataset Repository: this https URL ].

Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-09-10 in cs.CV).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2609.09895 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-16T10:02:34.340Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-09-10?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).