benchmarks evals
Every Comparable Model Lost Ground Across Repeated Encounters
EgoGears separated local perception from the harder job of binding evidence across independent route videos.
Summary
EgoGears separated local perception from the harder job of binding evidence across independent route videos.
EgoGears contains 567 single-video and 1,487 multi-video questions from 126 egocentric recordings of 39 outdoor routes; 531 questions require alignment across recordings. Among 20 model configurations tested comparably on both splits, every one performed worse on multi-video questions, with a mean decline of 22.5 percentage points. Controls indicate that additional video volume or recording boundaries alone do not explain the gap. The authors identify observation-evidence binding and ordered route-state tracking as the central tested bottlenecks.
Why it matters
EgoGears separated local perception from the harder job of binding evidence across independent route videos.
Limits and context
- Controls indicate that additional video volume or recording boundaries alone do not explain the gap.
Key claims
EgoGears separated local perception from the harder job of binding evidence across independent route videos.
Qualification: Controls indicate that additional video volume or recording boundaries alone do not explain the gap.
Evidence: source-2026-09-30-015
Sources
- arXiv preprint 2609.37938arXiv · primary research
Corrections
No corrections have been recorded for this story.