domains / families / research / naacl-reproducibility

naacl-reproducibility

separate

Use when hardening the reproducibility story of a NAACL submission — treating the Responsible NLP checklist as a binding contract, pinning model versions and API access dates, making multilingual evaluation re-runnable, and stating an honest release level instead of an aspirational one.

When to use

• Hardening NAACL submission reproducibility
• Auditing Responsible NLP checklist answers
• Pinning model versions and API access dates
• Making multilingual evaluation re-runnable
• Stating honest release levels
• Creating reproducibility manifests
• Logging per-language preprocessing
• Preparing for author response windows

When not to use

• General NLP paper writing
• Non-NAACL conference submissions
• Ethical AI guidelines unrelated to reproducibility
• Code optimization
• Data collection strategies

Skill page

Source