Viewing profile — upperhalfplane
upperhalfplane
HN member- Joined
- Sun, Feb 02, 2025, 3:16 PM UTC
- HN karma
- 13
- Public activity
- 10 items
- HN profile
- View on Hacker News ↗
About upperhalfplane
No profile information was provided.
Recent public activity
-
comment
Comment #47452420
My favorite geometric proof of an inequality is the one I read on Terry Tao's blog. Interestingly, it's not presented as a geometric proof, but it is very much one: if you have two…
-
story
Show HN: Xent Game: help GPT2 minimize its surprise
Hi there! We're Andrew and Clément and we decided to make a game to understand better how LLM process and compress information. So we made a sort of reverse prompting game with con…
- story
-
comment
Comment #44913939
TLDR: games are a good way to go.
-
comment
Comment #44913810
I don't quite think they cheat at math olympiads, but obviously there are blindspots for the unspectacular tasks. That being said, Mississippi is both a good and a bad question to …
-
comment
Comment #44887346
> It's basically always guessing / confabulating / hallucinating if you ask it an introspective question like that. You're absolutely right! This is the basis of this recent paper …
-
comment
Comment #44616733
100% agree about this too (also a professional mathematician). To mathematicians who have not been trained on such problems, these will typically look very hard, especially the mor…
-
comment
Comment #43117700
I don't think it's an order of magnitude below... though it's a bit hard to know for sure
-
comment
Comment #43115731
It looks like you can (there are some LLM responses out there, e.g. Sonnet 3.5). Not clear if they can be super good at this, though.
- story