Metraipolis › archive › dc_527db601df06

Visual Abstention in Unified Multimodal Models — arXiv:2610.07887

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.07887
License (as archived)
CC0-1.0
Archived
2026-10-07T11:21:17.985Z
SHA-256 of the archived text
7076bd5cb288511a00967f276da1ffad2bda63c147af8a0927145e4036f47827
Stable URL of this page
https://metraipolis.com/documento/dc_527db601df06

How to cite

Visual Abstention in Unified Multimodal Models — arXiv:2610.07887 — arXiv (en), https://arxiv.org/abs/2610.07887 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_527db601df06 (Metraipolis archive dc_527db601df06, sha256 7076bd5cb288511a00967f276da1ffad2bda63c147af8a0927145e4036f47827)
BibTeX
@misc{dc527db601df06,
  title = {Visual Abstention in Unified Multimodal Models — arXiv:2610.07887},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.07887},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_527db601df06; sha256 7076bd5cb288511a00967f276da1ffad2bda63c147af8a0927145e4036f47827},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.07887. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Unified multimodal models (UMMs) integrate understanding and generation, yet their generative behavior is rarely governed by what they understand about the task. We formalize visual abstention: when a requested visual transformation is impossible under the task's rules, the model should recognize that no valid solution exists, state this, and decline to generate. We introduce Draw-or-Decline (DoD), a benchmark of 1,050 feasible-infeasible request pairs across 7 task categories that jointly measures editing success and the refusal of infeasible requests. Evaluating 8 UMMs, we find that editing ability and abstention are distinct capabilities: even the strongest editor, at 68.4% editing accuracy, refuses only 0.4% of infeasible requests under ordinary instructions. Their reasoning shows why: the models rarely notice the conflict, and instead plan the edit as if the request were possible, often describing objects that are not in the image, or quietly change the request into one they can complete. Explicitly prompting these UMMs to report infeasibility increases textual refusals but reduces editing accuracy. We propose VisTA (Visual Transformation and Abstention), a training method that pairs feasible and infeasible examples so that a model judges feasibility before deciding whether to generate. We train VisTA-BAGEL to perform feasible edits and decline infeasible requests. Without any reminder, it refuses 93.0% of infeasible requests, up from 0.4% for the strongest editor, while falsely refusing only 0.8% of feasible ones. Unlike a reminder, this does not cost editing accuracy: VisTA-BAGEL completes 74.3% of feasible edits, more than any of the 8 evaluated UMMs.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2610.07887 — 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 2026-10-07T11:15:24.778Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).