StoryResearch
A paper written entirely by The AI Scientist-v2 passes peer review at an ICLR 2025 workshop
Sakana submitted three fully AI-generated papers to an ICLR 2025 workshop. One, reporting a negative result on a regularisation method, scored 6, 7 and 6, above the average acceptance bar. The team withdrew it before publication because the field has not decided how to handle AI-written papers.
- 3 papers submitted; 1 passed review with an average score of 6.33.
- No human edits: hypothesis, code, analysis and text were all AI-generated.
- Withdrawn before publication as part of the experiment's design.