emnlp-author-response
GitHub用于EMNLP ARR审稿期间撰写作者回复，指导按合理性或创新性分类处理意见，明确可补充内容边界，提供高价值回复结构模板，确保匿名性并针对AC和SAC优化表述。
Trigger Scenarios
Install
npx skills add brycewang-stanford/Awesome-Journal-Skills --skill emnlp-author-response -g -y
SKILL.md
Frontmatter
{
"name": "emnlp-author-response",
"description": "Use when drafting an EMNLP author response in the ACL Rolling Review discussion window — triaging replies by soundness versus excitement, deciding what evidence can be added mid-review, staying anonymous, and writing for the area chair's meta-review and the post-commitment Senior Area Chairs, not the reviewers alone."
}
EMNLP Author Response
Use this when ARR reviews land. In the May 2026 cycle the author response and author-reviewer discussion ran July 7-13, with meta-reviews released July 30 and EMNLP commitment due August 2 — a pipeline where the response is written once but read three times: by reviewers who may update, by the AC composing the meta-review, and by the conference SACs who see the whole record after commitment. Confirm the current window mechanics on the live ARR pages before drafting.
Triage by score axis
Sort every reviewer point by which score it threatens:
- Soundness attacks ("the baseline is mistuned", "no significance test", "the claim exceeds the tested languages") — answer with evidence or a scope concession. These are the points that move outcomes, because soundness anchors publishability.
- Excitement discounts ("incremental", "unsurprising") — answerable only by reframing what the finding changes for the field, briefly, once. Arguing taste at length reads as not having evidence.
- Misreadings — correct with a quote and a section number, tone-free. A factual correction the AC can verify in ten seconds is your highest-value paragraph.
What can and cannot be added
The response box takes text and, in ARR practice, a revised PDF may accompany a resubmission rather than a same-cycle response — do not build the plan around uploading new material unless the current cycle's instructions explicitly allow it. Within text:
| Addition | Safe? | Why |
|---|---|---|
| Numbers from runs completed during the window | Usually | Reviewers can weigh them; label as new |
| A significance test on already-reported results | Yes | It re-analyzes existing evidence |
| A promised camera-ready rewrite | Yes, if scoped | Commit to wording, not to new experiments |
| A brand-new experimental direction | No | Unreviewable in-window; save it for revise-and-resubmit |
| Links to external results pages | No | Anonymity and auditability both break |
Anatomy of a reply that moves a meta-review
R2.1 (baseline tuning, soundness): Correct that §5.2 did not state the search budget.
Both systems received identical 32-trial random search over the grid in App. C; we
will state this in §5.2. Under matched budgets the gap is 2.1 F1 (±0.4 over 5 seeds,
paired bootstrap p=0.003, already in Table 3).
R2.4 (only high-resource languages, scope): We agree and will retitle the claim to
the six tested languages. We note Swahili and Tamil results in App. E show the same
direction at lower magnitude — we will surface this in §7 rather than claim
generality we did not test.
The pattern: number every reply to a numbered concern; concede early where the reviewer is right; convert each concession into a specific, checkable edit; anchor every number to a location in the submitted record.
EMNLP-typical objections and their strongest answers
- "Results may reflect data contamination." Point to the contamination audit (overlap statistics, cutoff dates) in the submission. If none exists, run the overlap analysis during the window and report it — this objection does not age away.
- "No error analysis." If the paper has one, the reviewer missed it — cite the section. If it doesn't, a compact categorization of 50-100 sampled failures, added as in-window text, is feasible and persuasive.
- "Human evaluation lacks detail." Report annotator count, guidelines location, agreement statistic, and pay — items the Responsible NLP checklist already made you record.
- "Why not compare against
?" Either the comparison exists under another name (say so), or explain the confound that makes it uninformative, or run it if the window permits. "Out of scope" without a reason is read as "we would lose."
When reviewers contradict each other
Three-reviewer packets regularly split: R1 wants more languages, R3 thinks the language set is already too broad for the method's claims. Do not answer each in isolation — the AC will read both replies side by side and see you agreeing with everyone. Instead:
- Name the disagreement once, neutrally: "R1 and R3 read the language coverage in opposite directions."
- State your position with its evidence, and accept the cost with the other reviewer explicitly.
- Give the AC the synthesis you want the meta-review to adopt — resolving reviewer conflict is the AC's job, and a response that does the work credibly usually gets adopted wholesale.
A special case is the score-text mismatch: a review whose text is mild but whose overall recommendation is harsh (or vice versa). Respond to the text — it is the only part you can engage — and let the mismatch itself be visible to the AC without commenting on it.
Deciding not to fight
Some packets should not be argued with. If all three reviewers converge on a missing piece of evidence that cannot be produced in the window — a new annotation effort, a second domain, a human study — the strongest move is a short, gracious response that corrects factual errors only, followed by revise-and-resubmit into a later cycle where the same reviewers will see the gap actually closed. A maximal rebuttal that concedes nothing, against a unanimous packet, damages the record that the SACs will later read at commitment time.
Discussion-window conduct
- Reply early; a July 7 response can get a reviewer follow-up, a July 12 response cannot.
- Keep the whole response skimmable — ACs in a 17,000-submission cycle read dozens of these; a one-screen summary block on top ("three concerns, three fixes") earns goodwill.
- Stay anonymous end to end: no identity hints, no "as our prior work" phrasing, no personal repositories.
- Never write toward the SACs explicitly ("we urge acceptance") — they read the record for resolved substance, and advocacy is not substance.
Output format
[Concern map] <Rx.y -> soundness / excitement / misreading / taste>
[Reply drafts] <numbered, evidence-anchored, concession-explicit>
[New-in-window evidence] <what was computed, labeled as new>
[Camera-ready commitments] <specific edits promised>
[Residual risk after response] <what remains unresolved and why>
Version History
- 9f86f09 Current 2026-07-19 15:17
