NeurIPS Peer Review is Highly Random: 50% of Spotlights Were Rejected in RCT

NeelNanda5 · x · 2026-07-24

AI researcher Neel Nanda shared practical advice for researchers going through the NeurIPS rebuttal phase. He advised authors not to take the review results too personally or be overly defensive, emphasizing the inherent randomness (RNG) in academic peer review.

Citing data from a randomized controlled trial (RCT) conducted before the use of LLMs, he pointed out that 50% of NeurIPS Spotlight papers were rejected by an independent review committee. This highlights that review outcomes involve significant luck, and researchers should focus on engaging constructively with genuine feedback.

Original post →

More from Research

Research channel →