{"slug":"brycewang-stanford-colm-experiments","source_name":"brycewang-stanford/colm-experiments","name":"Brycewang Stanford/Colm Experiments","description":"Use when designing or auditing the empirical core of a COLM paper — contamination analysis for evaluation data, fair baselines under matched prompting and compute, pinned model versions and decoding parameters, uncertainty over runs and samples, scaling coverage, and honest reporting of API-model comparisons.","version":1,"lift":{"pass_rate_delta_pts":27.27,"pass_rate_pct":95.5,"total_cases":22,"passed_cases":21,"tokens_delta_pct":46.3,"turns_delta_pct":0,"verdict":"mixed","benchmark_model":"gemini-3.6-flash","grading_method":"judged","completed_at":"2026-08-31T04:39:55.073536+00:00"},"skill_score":0.9545,"benchmark_models":[{"model":"gemini-3.6-flash","headline":true,"delta_pts":27.27,"with_pass_pct":95.5,"without_pass_pct":68.2,"tokens_delta_pct":46.3,"turns_delta_pct":0,"total_cases":22,"cases_aggregated":22,"verdict":"mixed","never_hurt":true,"completed_at":"2026-08-31T04:39:55.073536+00:00","run_id":"5d8b5b32-bf54-414e-a2ca-056c440b63d1","latest_attempt":{"id":"5d8b5b32-bf54-414e-a2ca-056c440b63d1","runner_model":"gemini-3.6-flash","overall_verdict":"mixed","completed_at":"2026-08-31T04:39:55.073536+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":"e5ea4df40cc9af8f2db6b0754cd0dc0897e40dd5ee0ee9bcf747b0c27fb9c580","install_name":"brycewang-stanford-colm-experiments","version_id":"79a3979d-d7f9-4286-a8d8-aed0f8032fa8","body_sha256":"e5ea4df40cc9af8f2db6b0754cd0dc0897e40dd5ee0ee9bcf747b0c27fb9c580","render_hash":"0a476f447fef239ea0e81dbc796afadb2030795de62c76264caa24dee140cb3d","raw_url":"https://app.decimal.ai/s/brycewang-stanford-colm-experiments/SKILL.md","scorecard_url":"https://app.decimal.ai/skills/brycewang-stanford-colm-experiments"}