HellaSwag
Not yet importedCommonsense completion tasks that ask a model to choose the most plausible continuation of an everyday situation.
2 case previewschatSource termsversion data / head
Straightforward import. Static prompt-and-label evaluation with no runtime dependencies.
What it measures
HellaSwag evaluates grounded commonsense by pairing a short context with several candidate endings. Adversarial filtering makes incorrect choices plausible enough to defeat shallow matching.
The benchmark fits Trunchbull's basic multiple-choice contract and does not need tools or a sandbox. The import work is primarily provenance, prompt normalization, and exact label scoring.
not yet imported
Not yet imported. A first release needs the dataset pinned, contexts normalized, and answer ordering preserved.
Cases
2 previewed| id | case | runtime |
|---|---|---|
| hellaswag/activitynet | Activity continuationChoose the most plausible next step in a described human activity. | chat |
| hellaswag/wikihow | Procedure continuationSelect the ending that coherently completes a procedural context. | chat |
Tools
none requiredNo model tools. Cases are answered from the prompt and checked by the grader.