{"slug":"mkurman-lighteval","source_name":"mkurman/lighteval","name":"Mkurman/Lighteval","description":"All-in-one LLM evaluation toolkit (HuggingFace LightEval). 1000+ tasks with multi-backend support: Accelerate, vLLM, SGLang, Nanotron, TGI, LiteLLM, inference providers, and custom models. Sample-by-sample result exploration, custom task/metric creation. Used by HuggingFace's Leaderboard and Evals team. For pure GPT-style model eval, also consider lm-evaluation-harness.","version":1,"lift":{"pass_rate_delta_pts":52.17,"pass_rate_pct":91.3,"total_cases":23,"passed_cases":21,"tokens_delta_pct":31.5,"turns_delta_pct":0,"verdict":"mixed","benchmark_model":"gemini-3.6-flash","grading_method":"judged","completed_at":"2026-08-21T10:02:24.775307+00:00"},"skill_score":0.913,"benchmark_models":[{"model":"gemini-3.6-flash","headline":true,"delta_pts":52.17,"with_pass_pct":91.3,"without_pass_pct":39.1,"tokens_delta_pct":31.5,"turns_delta_pct":0,"total_cases":23,"cases_aggregated":22,"verdict":"mixed","never_hurt":true,"completed_at":"2026-08-21T10:02:24.775307+00:00","run_id":"41cd8d17-9b77-4fb3-97e7-7c9e2b88d6e1","version_number":1,"is_latest_version":true,"gate":null}],"trust":{"skill_safety":"passed","safety_status":"clean","intent_verdict":"safe","content_status":"clean","indexable":true},"license":"MIT license","install_count":0,"manifest_hash":"6acc757ed4a8ba9f196370fcf5fa24acf7293082fe61bd0810531a53a856c79b","raw_url":"https://app.decimal.ai/s/mkurman-lighteval/SKILL.md","scorecard_url":"https://app.decimal.ai/skills/mkurman-lighteval"}