Reviewer Guideline Design for Automated Peer Review

Research applicability

Experimental study of how reviewer-guideline design affects LLM-based automated peer review, reporting that official conference guidelines align most closely with human judgments.

Getting started

Read the arXiv paper (2607.22553).

Maintenance

Maintenance status not independently audited.

License

Not independently audited.

Workflow & human oversight

Resource-level classification is intentionally unspecified; instantiated workflows may differ. Humans set objectives, supply data and check outputs.

Workflow level is not an evidence score; higher is not better evidence.

Approach

Varies reviewer-guideline designs used for LLM-based automated reviewing and evaluates agreement with human judgments, concluding that evaluation criteria refined through conference practice serve as effective guidance for automated reviewing.

Limitations

Conference acceptance is recorded in the arXiv comment field; the venue version was not inspected. Experimental results are author-reported and not independently reproduced.

Reproducibility

Live source checks 2026-10-17: arXiv metadata/abstract and comment field. No model, code or claimed outcome reproduced.

Adoption requirements

Research paper; no tooling release inspected in checked sources.