Please confirm you are human
This browser or connection looks automated. Press and continuously hold the control for 3 seconds to enable Google-hosted web results and, when separately allowed, AI-assisted answers.
A successful check enables 100 search requests. Interactive access does not authorize scraping, systematic collection, or reuse of search output.
News
One attention head carries knight forks in a chess transformer, and here's a new toolkit that found it. — LessWrong
1+ hour, 45+ min ago (19+ words) Quick interp demo in colab: Localize knight forks to a single head in Maia-3 with logit-lens and per-head ablation. …...
Watch a chess transformer think — LessWrong
4+ week, 17+ hour ago (96+ words) I was poking around at a chessformer which mimics human play and made this fun companion app to visualize a lot of the internals of the engine. Check out how attention heads look, how the residual stream evolves, and play…...
Dev log: implementing new syntax in GHC — LessWrong
1+ mon, 2+ week ago (1292+ words) In September 2024, I started working on a patch to GHC. The story starts slightly before that. In July 2024 I wondered Could UndecidableInstances be per-instance?. It seemed like a fairly small improvement that I could possibly even implement myself. The answer…...
What is a game? — LessWrong
1+ mon, 4+ week ago (909+ words) Some people think games are defined by their rules; Others define them by the behaviors of the players. This can lead to some misunderstandings, so I think it could be helpful foreach type of player to understand the other perspective....
How to get better at chess (and everything else) — LessWrong
3+ mon, 5+ day ago (534+ words) I've been following chess grandmaster Avetik Grigoryan for his chess improvement tips for a while. He has a tonne of great stuff on his website. The…...
Stockfish is not a chess superintelligence (and it doesn't need to be) — LessWrong
4+ mon, 3+ day ago (427+ words) TLDR: I demonstrate Stockfish is not a chess superintelligence in the sense of understanding the game better than all humans in all situations. It still kicks our ass. In the same way, AI may end up not dominating us in…...
Generalisation isn't actually (that) important — LessWrong
4+ mon, 3+ day ago (374+ words) TLDR: I demonstrate Stockfish is not a chess superintelligence in the sense of understanding the game better than all humans in all situations. It still kicks our ass. In the same way, AI may end up not dominating us in…...
Interpreting Gradient Routing’s Scalable Oversight Experiment — LessWrong
4+ mon, 1+ week ago (999+ words) This post is mainly about two things. First, it’s about the importance of simple baselines. Second, it's about different ways of modeling Scalable Oversight, and what each way entails. Let's start with some background for the first theme. Gradient Routing…...
I'm starting a substack — LessWrong
4+ mon, 3+ week ago (5+ words) Discuss I'm starting a substack...
Rational Chess — LessWrong
5+ mon, 1+ week ago (540+ words) This post is intended to be understandable even if you have never played chess, but will probably not interest you if you don't like games like chess. TL,DR: The most fair version of chess would be Armageddon (if neither…...