Reasoning
EgoSchema (released LLoVi configurations)
Published LLoVi question-answering results on the 500-question public EgoSchema subset, retaining the available caption-based inputs and eleven model/prompt configurations.
4,500items
11subjects
No explicit license was found for the captured LLoVi output/data archives or EgoSchema annotations. Do not infer MIT permission from the previous builder.license
reasoningdomain
textmodality
item-level responses released
Saturation status: No
Response matrix
Loading response matrix…
Correct (1)Incorrect (0)Unobserved
Scale: 1 = correct · 0 = incorrect