WINOGRANDE: An Adversarial Winograd Schema Challenge at Scale

Active

Set of 273 expert-crafted pronoun resolution problems originally designed to be unsolvable for statistical models that rely on selectional preferences or word associations.

Publisher: Allen Institute for AI
Domain: Reasoning
License: mit
Published: May 2026
Notable for: Benchmark for evaluating Reasoning.
Canonical: github.com/UKGovernmentBEIS/inspect_evals/tree/main/src/inspect_evals/winogrande

Cite

Notes

Only stored in your browser.

Attribution

README: github.com/UKGovernmentBEIS/inspect_evals/blob/main/src/inspect_evals/winogrande/eval.yamlMIT

Attribution policy →

FAQ

What is WINOGRANDE: An Adversarial Winograd Schema Challenge at Scale?: Set of 273 expert-crafted pronoun resolution problems originally designed to be unsolvable for statistical models that rely on selectional preferences or word associations.
What license is WINOGRANDE: An Adversarial Winograd Schema Challenge at Scale under?: WINOGRANDE: An Adversarial Winograd Schema Challenge at Scale is available under mit.