Use when designing or auditing the experimental program of an EMNLP paper — baseline fairness and tuning budgets, dataset and language coverage matched to claims, contamination audits for pretrained and API models, significance testing and seed variance, human evaluation protocols, prompt sensitivity, and the error analysis EMNLP reviewers expect.
