Use when hardening an EMNLP paper's reproducibility record — answering the Responsible NLP checklist truthfully under desk-reject enforcement, pinning model versions and API snapshot dates, logging decoding parameters and prompts, documenting data licensing and annotation, and reporting compute so another lab could rerun the study.
