{"slug":"openlair-simpo-training","source_name":"openlair/simpo-training","name":"Openlair/Simpo Training","description":"Simple Preference Optimization for LLM alignment. Reference-free alternative to DPO with better performance (+6.4 points on AlpacaEval 2.0). No reference model needed, more efficient than DPO. Use for preference alignment when want simpler, faster training than DPO/PPO.","version":1,"lift":{"pass_rate_delta_pts":34.78,"pass_rate_pct":91.3,"total_cases":23,"passed_cases":21,"tokens_delta_pct":33.3,"turns_delta_pct":0,"verdict":"mixed","benchmark_model":"gemini-3.6-flash","grading_method":"judged","completed_at":"2026-08-07T20:00:22.811043+00:00"},"skill_score":0.913,"benchmark_models":[{"model":"gemini-3.6-flash","headline":true,"delta_pts":34.78,"with_pass_pct":91.3,"without_pass_pct":56.5,"tokens_delta_pct":33.3,"turns_delta_pct":0,"total_cases":23,"cases_aggregated":21,"verdict":"mixed","never_hurt":true,"completed_at":"2026-08-07T20:00:22.811043+00:00","run_id":"77bffdaf-f5a1-48da-b1bb-025be1e77f76","latest_attempt":{"id":"77bffdaf-f5a1-48da-b1bb-025be1e77f76","runner_model":"gemini-3.6-flash","overall_verdict":"mixed","completed_at":"2026-08-07T20:00:22.811043+00:00","gate":null},"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","install_count":0,"total_uses":0,"total_users":0,"manifest_hash":"91990954d999fa3af6e231efd395ce1d76cd659f369816bd8c82312c6717f1d9","install_name":"openlair-simpo-training","version_id":"d0254ecc-8864-481a-8557-5809c1cee3ca","body_sha256":"91990954d999fa3af6e231efd395ce1d76cd659f369816bd8c82312c6717f1d9","render_hash":"cf09d3f6218492200e83759e3d5c78f271b30616396445218103caa3d29d803a","raw_url":"https://app.decimal.ai/s/openlair-simpo-training/SKILL.md","scorecard_url":"https://app.decimal.ai/skills/openlair-simpo-training"}