domains / families / science / eacl-experiments

eacl-experiments

separate

Use when designing or auditing the empirical evidence for an EACL paper, covering tuned and LLM baselines, multilingual breadth matched to the claim, significance and variance floors, human-evaluation agreement, data-contamination controls, ablations, and error taxonomies, so that every stated result is measured rather than asserted.

When to use

• designing empirical evidence for EACL papers
• auditing empirical evidence for EACL papers
• defining tuned and LLM baselines
• matching multilingual breadth to claims
• setting significance and variance floors
• reporting human-evaluation agreement
• implementing data-contamination controls
• creating ablations and error taxonomies

When not to use

• leaderboard maximalism
• unscoped claims
• strawman baselines
• unreported LLM baselines
• lone run reporting
• unmeasured human evaluations
• unquantified error analysis

Skill page

Source