OpenParable
Skip to content
← Back to skills

Neurips Supplementary

Documents
neurips-supplementary

Neurips Supplementary is an Agent Skill for documents published in the brycewang-stanford/Awesome-Journal-Skills repository on GitHub: a SKILL.md instruction file an AI agent can load to do this task.

Use when deciding what NeurIPS material belongs in the main PDF body, the references, text appendices, the mandatory Paper Checklist, the separate code/data ZIP, or an external public artifact, and how to keep every supplemental component anonymous under double-blind review.

View on GitHub ↗
QUICK START

How to use this skill

Bring this guide into your coding agent with a prompt tailored to the tool you use.

  1. Open your project in Codex.
  2. Copy and paste the prompt below.
  3. Review files and risks before approving any action.
Prompt to paste
Inspect the Agent Skill at https://github.com/brycewang-stanford/Awesome-Journal-Skills/blob/HEAD/NeurIPS-Skills/skills/neurips-supplementary/SKILL.md and all supporting files. Report provenance, permissions, scripts, dependencies, and security risks. Propose an exact installation into .agents/skills/neurips-supplementary, but wait for my approval before writing files or running scripts.

Copying this prompt does not install or run the skill. Review third-party files before use.

NeurIPS Supplementary

Use this skill to prevent two common NeurIPS failures: hiding the main contribution in the appendix and overloading the main paper with details that belong in supplemental material.

Placement rules

  • Main paper: problem, contribution, method, essential theory, core experiments, limitations, and enough detail for reviewers to judge correctness.
  • References: complete, verified citations; do not rely on AI-generated references without manual verification.
  • Text appendices: proofs, derivations, additional ablations, dataset details, model cards, extended limitations, and supporting analyses.
  • Checklist: mandatory and placed according to current template instructions.
  • Separate ZIP: code, data, scripts, configuration files, logs, lightweight demo assets, or other review artifacts when allowed by current rules.

NeurIPS-specific guardrails

  • The supplement does not rescue an unclear main claim. Reviewers are not required to read every supplemental detail.
  • Supplementary material must be anonymized during double-blind review.
  • Links must preserve anonymous browsing and comply with current policy.
  • Code and data that support central claims should include exact commands and environment details, not just a repository dump.

Placement decision table

NeurIPS assembles the final PDF from distinct zones plus a separate review ZIP. Misplacing material is a recurring weakness: a proof that a reviewer cannot find, or a contribution buried where it does not count. Map each item to exactly one home before submission; reopen the current template for the authoritative zone list.

MaterialHomeReason it goes there
Central claim, headline resultMain bodyreviewers judge the contribution here
Full proofs and derivationsText appendixbody states the theorem; appendix carries the proof
Extra ablations, seeds, scalingText appendixsupports but does not replace core experiments
Dataset/model cards, datasheetsText appendix + ZIPprovenance reviewers expect, anonymized
Mandatory Paper ChecklistTemplate-specified slotdesk-reject risk if missing or misplaced
Runnable code, configs, logsSeparate ZIPreview-only artifact, not body pages
Broader-impact discussionMain bodya NeurIPS norm reviewers look for, not appendix-only

Datasets & Benchmarks supplement note

For a Datasets & Benchmarks track submission, the datasheet, license, intended-use statement, and hosting/maintenance plan are not optional appendix garnish; they are part of what reviewers score. Keep them anonymized for review and stage a de-anonymized public version for release. Hedge the exact required documentation to the current year's track call.

Worked vignette: a benchmark paper

Authors put their 40-page datasheet, croissant metadata, and loading code only in the appendix and leave the main body to leaderboard tables. A NeurIPS reviewer cannot judge dataset quality from tables alone. The fix: lift provenance, licensing, consent, and intended-use sentences into the main body so the contribution is visible; keep the full datasheet and loaders in the appendix and ZIP; and add one navigation sentence telling reviewers exactly where the documentation lives.

Output format

[Main-paper must stay] <items>
[Move to appendix] <items>
[Move to ZIP/artifact] <items>
[Anonymity fixes] <items>
[Reviewer navigation] <one sentence telling reviewers where to find support>