Reproduce “Comparative Evaluation of LLM-Based Review Paradigms for Academic Paper Quality Assessment” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Fan Huang, Hongyu Zhou

The paper tests whether LLM scoring methods can help triage conference submissions by comparing five paradigms across models and inputs. On 232 papers, GPT-4o-mini Direct and Elo have similar rank association with reviewer scores, but differ at a fixed cutoff. In 11,513 submissions, a base-rate-matched diagnostic gives Direct higher Cohen’s κ than Elo, without establishing a deployable gate.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.