HellaSwag

Official

Sentence completion benchmark testing grounded commonsense inference. Models must pick the most plausible continuation of an activity description.

Category: reasoningRunner: lm-eval-harnessVersion: v1.0Submitted by: Community

Eval Details

Scoring

Exact Match

Aggregation

Mean

Direction

Higher is better

Tasks

1 task

Seed: 42FewShot: 10

Task	Dataset	Weight	Shots	Max Tokens
HellaSwag hellaswag	Rowan/hellaswag / validation	1	10-shot	—

No approved results yet. Submit a run via the API.