VALSE

anchor

Findings

• IC-1389Video-language models do not significantly outperform image-language models on temporal reasoning tasks in VILMA [context]
• IC-456VLM decoders achieve near-random accuracy on VALSE image-sentence alignment while pairwise accuracy is much higher, indicating reliance on linguistic priors [builds-on]
• IC-456VLM decoders achieve near-random accuracy on VALSE image-sentence alignment while pairwise accuracy is much higher, indicating reliance on linguistic priors [eval]
• IC-458Most VLM decoders show negative CC-SHAP on VALSE multiple-choice, indicating their explanations are less self-consistent than their answers, driven by a shift from text-dominant to image-dominant processing [eval]