Stable Scores, Unstable Answers: Frame Phase and Option Order in Video Multiple-Choice Evaluation
Video-language models are ranked by multiple-choice accuracy on frames from a uniform grid.
ProofPaper ↗
Key points
- The grid has two parameters, a rate and a phase, and benchmarks report only the rate.
- The phase moves answers: two deployed samplers differing only by a half-step phase offset answer 23.6% of questions differently while scoring within a point, and across four releases from two families shifting only the phase changes roughly one answer in five after controlling option order.
- PHASEFUSION decodes three offset grids and averages the option posteriors.
- Option order, which changes only the presentation, is flagged instead by a one-pass answer margin.
Sources (1)
- [1]Stable Scores, Unstable Answers: Frame Phase and Option Order in Video Multiple-Choice EvaluationarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 6, 04:36 PM
Video-language models are ranked by multiple-choice accuracy on frames from a uniform grid.
The grid has two parameters, a rate and a phase, and benchmarks report only the rate.
Extractive summary: sentences quoted from the sources.