{"slug":"dataset-labeling-guidelines","source_name":"dataset-labeling-guidelines","name":"Write Labeling Guidelines That Produce Consistent Labels","description":"Turns a labeling or annotation task into a guideline different annotators can apply the same way — label definitions with explicit in/out boundaries, a worked borderline example for each label, a tie-break rule for when two labels fit or a case is unclear, canonical gold examples, and an agreement check across annotators. Use when someone is setting up a labeling, tagging, annotation, or coding task for a dataset and wants the instructions or rubric to hand to labelers. Do NOT use for building or training the classifier itself, for cleaning or deduplicating already-labeled data, or for writing end-user content policy rather than annotator instructions.","version":1,"lift":{"pass_rate_delta_pts":null,"pass_rate_pct":53.8,"total_cases":13,"passed_cases":7,"tokens_delta_pct":542.6,"turns_delta_pct":0,"verdict":"mixed","benchmark_model":"gemini-3.5-flash","grading_method":"judged","completed_at":"2026-07-08T21:58:57.679679+00:00"},"skill_score":null,"benchmark_models":[{"model":"gemini-3.5-flash","headline":false,"delta_pts":null,"with_pass_pct":null,"without_pass_pct":null,"tokens_delta_pct":542.6,"turns_delta_pct":0,"total_cases":13,"cases_aggregated":13,"verdict":"mixed","never_hurt":true,"completed_at":"2026-07-08T21:58:57.679679+00:00","run_id":"2e5192c6-bb82-4529-8788-a2d138971305","version_number":1,"is_latest_version":true,"gate":"not_comparable"}],"trust":{"skill_safety":"passed","safety_status":"clean","intent_verdict":"safe","content_status":"clean","indexable":false},"license":null,"install_count":0,"manifest_hash":"9382e85ac8f6f1126f2bf3b2cad6d0b77c75a2896e822dd18e78f58dd69d9da5","raw_url":"https://app.decimal.ai/s/dataset-labeling-guidelines/SKILL.md","scorecard_url":"https://app.decimal.ai/skills/dataset-labeling-guidelines"}