Not yet importedNo sandbox required

WinoGrande

Upstream publisher Allen Institute for AIallenai/winogrande

Large-scale pronoun-resolution problems that require commonsense reasoning to resolve an ambiguous reference.

commonsensecoreferencechat

PROVENANCEUPSTREAM RECORD

Upstream source: allenai/winogrande
Trunchbull port: Not packaged
Version: XL
Release digest: Not compiled
License: Dataset terms

Source commit: Upstream head not pinnedPublished Indexed July 30, 2026

WHAT THIS PROJECT MEASURES

README

WinoGrande presents sentences with a blank and two candidate referents. The correct answer depends on commonsense rather than local grammar alone.

The benchmark maps cleanly to Trunchbull's exact-choice evaluation path. No model tools, code execution, or sandbox environment are needed.

EXECUTION CONTRACT

Static two-choice evaluation in the standard chat lane.

TRUNCHBULL AVAILABILITY

Import status

Not yet imported

Not yet imported. A faithful release needs the selected split pinned and answer positions rendered without leaking labels.

Port this benchmark

1 CASES

Evaluation cases

winogrande/xl

Commonsense coreference

Chat lane

Resolve an ambiguous reference using the surrounding situation.

0 REQUESTED TOOLS

Tools

No model tools required.

Evaluation happens through the prompt and grader contract for this release.

MACHINE-READABLE PROVENANCE

Registry record

{
  "name": "WinoGrande",
  "catalogOwner": null,
  "source": {
    "url": "https://github.com/allenai/winogrande",
    "version": "WinoGrande XL",
    "commit": "Upstream head not pinned",
    "license": "Dataset terms"
  },
  "port": null,
  "release": {
    "digest": "Not compiled",
    "lane": "chat",
    "sandboxRequired": false
  },
  "cases": [
    "winogrande/xl"
  ],
  "tools": []
}

BUILD ON THIS PROJECT