Reproduce “Leaderboard Hacking: Preference-Based Model Evaluations are Vulnerable to Manipulation” 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

Eve Fleisig, Joachim Baumann, Dylan Hadfield-Menell, Dirk Hovy

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