Papers
arxiv:2611.25776

Suppressing the Shortcut Hidden Behind the Gains of VideoLLM Fine-Tuning

Published on Nov 19
· Submitted by
Youngwoo Shin
on Nov 20
Authors:
,

Abstract

VETO suppresses shortcut learning in VideoLLMs by projecting supervised fine-tuning updates onto the orthogonal complement of the text-only activation subspace, improving out-of-distribution video reasoning.

Supervised fine-tuning (SFT) is the stage in which Video Large Language Models (VideoLLMs) advance their capability to reason over video. However, the VideoLLM also learns to answer from textual regularities of its training data, a shortcut that bypasses the video. The regularities behind the shortcut are hard to define at the data level, and the shortcut forms regardless of training epoch, data size, or backbone. The shortcut is hidden behind a rise in accuracy within the training distribution, yet it is not required for the rise and even costs SFT the gain the video offers under distribution shift. We therefore propose Video-Evidence Training via Orthogonal projection (VETO), which intervenes on the fine-tuning update where the shortcut is written rather than on the data. VETO calibrates the subspace that the text alone occupies in the activations of the base model and projects the update onto its orthogonal complement. The VideoLLM is thus left to ground its answer in the full input, with the shortcut suppressed as it forms and never defined. Across seven VideoLLMs, VETO preserves the gain of SFT on the training distribution and suppresses the shortcut. On six benchmarks under shift it gains more than SFT and answers from the full input rather than the shortcut. Code is available at https://github.com/Youngwoo-git/VETO.

Community

Paper author Paper submitter

SFT raises the accuracy of VideoLLMs, yet behind this rise the model increasingly relies on textual regularities of the training data, a shortcut that bypasses the video. VETO suppresses this shortcut by projecting the fine-tuning update onto the orthogonal complement of the subspace the text alone occupies, so the model keeps the gains of SFT while grounding its answer in both the video and the text.

Sign up or log in to comment

Get this paper in your agent:

hf papers read 2611.25776
Don't have the latest CLI?
curl -LsSf https://hf.co/cli/install.sh | bash

Models citing this paper 0

No model linking this paper

Cite arxiv.org/abs/2611.25776 in a model README.md to link it from this page.

Datasets citing this paper 0

No dataset linking this paper

Cite arxiv.org/abs/2611.25776 in a dataset README.md to link it from this page.

Spaces citing this paper 0

No Space linking this paper

Cite arxiv.org/abs/2611.25776 in a Space README.md to link it from this page.

Collections including this paper 0

No Collection including this paper

Add this paper to a collection to link it from this page.