Skip to main content
Datasetbenchmarksv1.1

HellaSwag Dataset

by University of Washington · open-source · Last verified 2026-03-17

HellaSwag is an adversarially filtered commonsense NLI benchmark where models must pick the most plausible sentence completion from 4 options. Humans score 95%+ while early LLMs struggled below 50%, making it a robust test of grounded language understanding and commonsense reasoning.

https://huggingface.co/datasets/Rowan/hellaswag
B+
B+Good
Adoption: A+Quality: AFreshness: B+Citations: A+Engagement: F

Specifications

License
MIT
Pricing
open-source
Capabilities
commonsense-evaluation, sentence-completion-benchmark
Integrations
huggingface-datasets, lm-eval-harness
Use Cases
model-evaluation, commonsense-reasoning, benchmarking
API Available
No
Tags
benchmark, commonsense, sentence-completion, adversarial, grounding
Added
2026-03-17
Completeness
100%

Index Score

77
Adoption
91
Quality
88
Freshness
70
Citations
92
Engagement
0

Put AI to work for your business

Deploy this dataset alongside autonomous AaaS agents that handle tasks end-to-end — no babysitting required.

Explore the full AI ecosystem on Agents as a Service