{
 "id": "richardyoung/code-safety-benchmark-results",
 "owner": "richardyoung",
 "name": "code-safety-benchmark-results",
 "downloads": 9,
 "likes": 0,
 "last_modified": "2026-04-21T22:28:45.000Z",
 "stated_licence": "other",
 "stated_from": "tag",
 "licence_name": "research-results-private-during-review",
 "licence_link": "https://github.com/ricyoung/code-safety-benchmark/blob/main/LICENSE",
 "file_licence": null,
 "licence_files": [],
 "sources": [
  {
   "id": "richardyoung/code-safety-prompt-bank",
   "licence": null,
   "found": false,
   "evidence": "https://huggingface.co/datasets/richardyoung/code-safety-prompt-bank"
  }
 ],
 "gated": "auto",
 "tasks": [
  "text-classification"
 ],
 "citation": "@article{young2026codebench,\n  title   = {Is Code Safety a Separate Direction? A 13-Model Behavioral Benchmark of Malicious-Code Refusal in Coding-Specialized LLMs},\n  author  = {Young, Richard J. and Moody, Gregory D.},\n  journal = {arXiv preprint arXiv:2604.XXXXX},\n  year    = {2026}\n}",
 "signal": "unclear",
 "reason": "The checks could not identify a recognised licence ID in the dataset's metadata.",
 "revision": "b37a08195885cdebc03e7b6b4c6fe29e573a0631",
 "stated_evidence": "https://huggingface.co/datasets/richardyoung/code-safety-benchmark-results/blob/b37a08195885cdebc03e7b6b4c6fe29e573a0631/README.md",
 "file_evidence": "https://huggingface.co/datasets/richardyoung/code-safety-benchmark-results/tree/b37a08195885cdebc03e7b6b4c6fe29e573a0631",
 "card_nc": false
}