GPT-5.6 disproves 20-year-old Benjamini-Hochberg statistical conjecture
The model constructed a counterexample demonstrating FDR control failure.
My impression https://web.stanford.edu/~lindrew/stats300C.pdf based on 300c (emmanuel's class , which I think @EdgarDobriban and I took together back in 2013 ) is that if you do BH-q , we kind of expect that you control at \approx q but probably >q. FDR experts feel free too…
Reactions
4 postsHP 
Hieu Pham@hyhieu226not even news 🥹 https://twitter.com/polynoamial/status/2077762676932165996
- likes: 42
- replies: 1
- bookmarks: 0
- reposts: 0
NB Noam Brown@polynoamial2023: LLMs struggle with 4th grade word problems 2024: LLMs can do high school math 2025: LLMs get a gold medal at the IMO Now, GPT-5.6 solves famous frontier math/stat questions. The IMO is today and 5.6 one-shotting a perfect score isn't even news. Where will we be next…
- likes: 2.6K
- replies: 80
- bookmarks: 547
- reposts: 252
AK SM Swaroop Mishra@Swarooprm7@polynoamial 2024: LLM got IMO silver 🙂
- likes: 2
- replies: 0
- bookmarks: 0
- reposts: 0