Agent Skills
› brycewang-stanford/Awesome-Journal-Skills
› eacl-supplementary
eacl-supplementary
GitHub指导EACL论文内容分配，确保决策关键信息位于正文。强调局限性章节为必读部分，附录与匿名归档仅作补充。提供放置层级、自包含测试及常见错误检查，保障评审体验。
Trigger Scenarios
整理EACL论文结构
安排补充材料位置
检查内容是否自包含
处理匿名归档
Install
npx skills add brycewang-stanford/Awesome-Journal-Skills --skill eacl-supplementary -g -y
SKILL.md
Frontmatter
{
"name": "eacl-supplementary",
"description": "Use when organizing an EACL paper's body, mandatory Limitations section, appendices, and anonymized supplementary archive so that no decision-critical content hides where ACL Rolling Review reviewers are not obliged to read, and so the content-page budget carries the whole argument on its own."
}
EACL Supplementary Material
Use this to decide what goes where across an EACL submission: the content pages, the Limitations section, the appendices, and the anonymized supplementary archive. The governing rule in ARR is that reviewers must read the content pages but are not obliged to read appendices or the archive — so anything decision-critical that lives outside the body is a self-inflicted risk. Reopen the current ARR CFP before finalizing structure.
The placement hierarchy
| Location | Reviewer obligation | Put here |
|---|---|---|
| Content pages (8 long / 4 short) | Must read | The claim, method, key results, the argument |
| Limitations (after conclusion, uncounted) | Must read | Honest scope, failure modes, threats to validity |
| References (unlimited) | As needed | Citations |
| Appendices (in-PDF, after refs) | May skip | Proof detail, extra tables, full hyperparameters |
| Anonymized archive (OpenReview) | May skip | Code, data, prompts, model outputs |
The self-containedness test
- Read only the content pages and the Limitations section. Can a reviewer accept or reject on that alone? If a core mechanism, the main baseline, or the headline result lives only in an appendix, move it into the body.
- "See Appendix X" is fine for detail, never for the argument.
Limitations is body, not backmatter
- The Limitations section is mandatory and read; treat it as part of the argument, not a disclaimer. Name the languages you did not cover, the settings you did not test, and the claims you cannot support — reviewers reward this and it is explicitly protected.
- Do not smuggle results into Limitations to dodge the page limit; that is transparent and counterproductive.
Appendix organization
Appendix A: full hyperparameters + search spaces
Appendix B: additional / per-language results tables
Appendix C: annotation guidelines + agreement
Appendix D: prompt templates + decoding settings (verbatim)
Appendix E: additional error analysis / examples
Label appendices so the body can point precisely, and keep each reproducible from the archive.
Anonymized archive hygiene
- The OpenReview supplement must be fully de-identified (see
eacl-artifact-evaluation): no names in paths, headers, or git history. - Keep the archive self-contained — a reviewer who does open it should be able to reproduce a reported table without hunting for external files.
Common misplacements to catch
- Core baseline hidden in an appendix — reviewers may never see it; the comparison then reads as missing.
- Key multilingual result only in the archive — an EACL reviewer will not go digging; surface at least a summary table in the body.
- Ethics/data-governance detail buried — for sensitive or lower-resource data, keep the intended-use and provenance statement visible.
Output format
[Placement audit] Pass / Needs moves
[Decision-critical content outside body] <list, with where it should go>
[Limitations quality] <substantive / ritual>
[Appendix map] <A-E contents>
[Archive hygiene] <anonymized + self-contained: pass/fail>
Version History
- 9f86f09 Current 2026-07-19 15:15
