界面已本地化;技能介绍与说明文档按所选语言由 AI 自动翻译,代码、命令、路径和专业术语保留原文,以原作者文档为准。
探索技能
数据与研究 · SKILL.md

neurips-camera-ready

neurips-camera-ready

【数据与研究】Use when preparing accepted NeurIPS papers for camera-ready upload, de-anonymization, final checklist, code/data release, OpenReview metadata, and post-acceptance obligations.

独立 GitHub 快照 · 自动分类 · 以原作者文档为准

技能说明

GitHub

name: neurips-camera-ready description: Use when preparing accepted NeurIPS papers for camera-ready upload, de-anonymization, final checklist, code/data release, OpenReview metadata, and post-acceptance obligations.

NeurIPS Camera-Ready

Use this skill only after acceptance. Camera-ready work is not just formatting: it is where anonymous review artifacts become public scientific records.

Official items to reopen

  • Current camera-ready instructions in OpenReview and the yearly NeurIPS handbook.
  • Final template and style file.
  • Final page, file-size, checklist, appendix, and upload rules.
  • Proceedings license, author metadata, presenter/registration requirements, and any presentation or poster instructions.

2026 baseline facts

The 2026 handbook states that accepted papers get one additional content page for camera-ready and that camera-ready upload happens through the OpenReview paper page by selecting the camera-ready version. It also says the final PDF combines paper pages, references, text appendices, and the paper checklist. Treat these as 2026 facts, not permanent rules.

Camera-ready workflow

  1. Resolve reviewer and AC issues that are allowed in camera-ready without changing the accepted contribution beyond recognition.
  2. De-anonymize authors, acknowledgments, repositories, datasets, model cards, code comments, and supplementary artifacts.
  3. Update the checklist so every answer reflects the final paper and public artifacts.
  4. Replace anonymous review links with stable public links, preferably with archived code and persistent dataset identifiers where appropriate.
  5. Verify that all figures, tables, citations, appendices, and artifact instructions match the final claims.
  6. Run a final integrity pass for hallucinated citations, broken links, hidden author metadata, and accidental private data.

Output format

[Camera-ready status] Ready / Needs fixes
[Official instructions checked] <URLs or missing items>
[Allowed paper changes] <summary>
[Public artifact tasks] <code/data/model/repo/archive>
[Final integrity risks] <citation/anonymity/license/privacy/checklist>
我的收藏