SOURCE-LINKED INTELLIGENCE

Does Video Memory Use What It Retrieves? A Causal Audit of Memory Specificity

arXiv · AI, language, vision and robotics · article · Sep 10, 2026 · UTC

Video models increasingly use memory to preserve information over long sequences, with the assumption that gains come from retrieving and using the correct past content. Standard memory ablations test whether memory helps, but not whether the retrieved content is responsible. We test this directly with read-time memory substitution, which replaces the consumed memory value while leaving the rest of the computation unchanged. This separates memory benefit from memory specificity, the extent to which the gain depends on retrieved content. Across frozen video world models, identity-free controls

Read original source ↗ Open in workspace

recordType

paper

region

Global