Datasets:
Languages:
English
Size:
n<1K
ArXiv:
Tags:
physical-reasoning
benchmark
vision-language-models
physics-simulation
minimal-pairs
evaluation
License:
Ctrl+K
Harness update: Antigravity CLI backend (tools denied, web search forbidden, transcript audit), Claude run tags, timeout fix; card documents the Antigravity safeguards (part 2)
1bdbac8 verified