I made a rating leaderboard by making papers fight each other. Helping overlooked papers get the recognition they deserve. How it works: papers are matched by rating, AI scores them on soundness, originality, experiments, clarity, and impact, then picks a winner. After thousands of battles, the leaderboard shows which papers hold up. 61K+ papers across NeurIPS, ICML, ICLR, ACL, EMNLP, NAACL. Full reasoning shown for every match. You can rebut the AI's judgment if you disagree. Random matches are free, no signup.