Author

J. Lee

h-index 1 View on Semantic Scholar

We have 1 of 9 papers

We haven’t gathered this author’s papers yet. Follow them and we’ll fetch their work.

Not the right person? Other researchers publish under this name.

Jul 2026

Accuracy Without Grounding: Diagnosing Visual Dependency Dissociation in Video LLM Benchmarks

The Visual Dependency Gap (VDG) is introduced, the difference in per-question correctness between original-video and black-screen conditions, which motivate VDG as a standard audit for whether video benchmarks measure visually grounded capability.

J. Lee · arXiv.org · 0 citations