AdamYedidia

Ban AI Su­per­in­tel­li­gence Form Let­ter

• AdamYedidia 1 Sep 2026 14:09 UTC

  11 points

  2

  in reply to: Linch’s comment on: Linch’s Shortform

  roon answered this on X:

  “i think answer to (1) [this question] is anodyne, afaict someone reclaimed a bunch of compute to run a different set of evaluations, thus interrupting many of the original evals”

  https://​​x.com/​​tszzl/​​status/​​2094149720251973899

  What links here?

  • Linch's comment on Linch’s Shortform by Linch (2 Sep 2026 7:00 UTC; 6 points)

AdamYe­di­dia’s Shortform

• AdamYedidia 1 Sep 2026 13:59 UTC

  9 points

  0

  on: AdamYe­di­dia’s Shortform

  I often see people say that knowledge of the HuggingFace hack is extremely low among the general US population, say 3%. Have there been polling/​surveys on this question, to anyone’s knowledge?

  Cheap billboards on random low-demand highways run at like half a cent per impression or less, is my sense. Which would mean buying one impression per person in the US would be about $2m. If someone were to stand up an attempt to buy cheap billboards telling people to google the HuggingFace hack and call their congresspeople encouraging them to shut down AI development, I would definitely donate.

• AdamYedidia 3 May 2026 3:06 UTC

  4 points

  2

  in reply to: Linch’s comment on: Linch’s Shortform

  Yep, it can tell it’s you!

• AdamYedidia 2 May 2026 18:11 UTC

  5 points

  0

  in reply to: Linch’s comment on: Linch’s Shortform

  Hit me with an unpublished blog post and I’ll run it through for you!

  What links here?

  • Linch's comment on Linch’s Shortform by Linch (5 May 2026 17:38 UTC; 4 points)

• AdamYedidia 25 Feb 2024 9:48 UTC

  8 points

  0

  on: Balanc­ing Games

  In Drawback Chess, each player gets a hidden random drawback, and the drawbacks themselves have ELOs (just like the players). As players’ ratings converge, they’ll end up winning about half the time, since they’ll get a less stringent drawback than their opponent’s.

  The game is pretty different from ordinary chess, and has a heavy dose of hidden information, but it’s a modern example of fluid handicaps in the context of chess.

• AdamYedidia 4 Nov 2023 21:55 UTC

  5 points

  0

  on: De­cep­tion Chess: Game #1

  (I was one of the two dishonest advisors)

  Re: the Kh1 thing, one interesting thing that I noticed was that I suggested Kh1, and it immediately went over very poorly, with both other advisors and player A all saying it seemed like a terrible move to them. But I didn’t really feel like I could back down from it, in the absence of a specific tactical refutation—an actual honest advisor wouldn’t be convinced by the two dishonest advisors saying their move was terrible, nor would they put much weight on player A’s judgment. So I stuck to my guns on it, and eventually it became kind of a meme.

  I don’t think it made a huge difference, since I think player A already had almost no trust in me by that point. But it’s sort of an interesting phenomenon where as a dishonest player, you can’t ever really back down from a suggested bad move that’s only bad on positional grounds. What kind of honest advisor would be “convinceable” by players they know to be dishonest?

De­cep­tion Chess: Game #1

• AdamYedidia 25 Oct 2023 20:29 UTC

  2 points

  0

  on: Ly­ing to chess play­ers for alignment

  I’d be excited to play as any of the roles. I’m around 1700 on lichess. Happy with any time control, including correspondence. I’m generally free between 5pm and 11pm ET every day.

• AdamYedidia 25 Oct 2023 20:25 UTC

  1 point

  0

  in reply to: GoteNoSente’s comment on: Chess as a case study in hid­den ca­pa­bil­ities in ChatGPT

  Oh wow, that is really funny. GPT-4′s greatest weakness: the Bongcloud.

• AdamYedidia 5 Oct 2023 22:31 UTC

  2 points

  0

  in reply to: Sheikh Abdur Raheem Ali’s comment on: New Tool: the Resi­d­ual Stream Viewer

  Sure thing—I just added the MIT license.

• AdamYedidia 2 Oct 2023 20:21 UTC

  2 points

  0

  in reply to: Sheikh Abdur Raheem Ali’s comment on: New Tool: the Resi­d­ual Stream Viewer

  Uhh, I don’t think I did anything special to make it open source, so maybe not in a technical sense (I don’t know how that stuff works), but you’re totally welcome to use it and build on it. The code is available here:

  https://​​github.com/​​adamyedidia/​​resid_viewer

New Tool: the Resi­d­ual Stream Viewer

• AdamYedidia 30 Sep 2023 20:25 UTC

  2 points

  0

  in reply to: GoteNoSente’s comment on: Chess as a case study in hid­den ca­pa­bil­ities in ChatGPT

  Good lord, I just played three games against it and it beat me in all three. None of the games were particularly close. That’s really something. Thanks to whoever made that parrotchess website!

• AdamYedidia 19 Aug 2023 17:49 UTC

  4 points

  1

  in reply to: dr_s’s comment on: Chess as a case study in hid­den ca­pa­bil­ities in ChatGPT

  I don’t think it’s a question of the context window—the same thing happens if you just start anew with the original “magic prompt” and the whole current score. And the current score is alone is short, at most ~100 tokens—easily enough to fit in the context window of even a much smaller model.

  In my experience, also, FEN doesn’t tend to help—see my other comment.

• AdamYedidia 19 Aug 2023 17:47 UTC

  4 points

  3

  in reply to: tailcalled’s comment on: Chess as a case study in hid­den ca­pa­bil­ities in ChatGPT

  It’s a good thought, and I had the same one a while ago, but I think dr_s is right here; FEN isn’t helpful to GPT-3.5 because it hasn’t seen many FENs in its training, and it just tends to bungle it.

  Lichess study, ChatGPT conversation link

  GPT-3.5 has trouble from the start maintaining a correct FEN, and makes its first illegal move on move 7, and starts making many illegal moves around move 13.

Chess as a case study in hid­den ca­pa­bil­ities in ChatGPT

• AdamYedidia 14 Aug 2023 22:41 UTC

  1 point

  0

  in reply to: RGRGRG’s comment on: The po­si­tional em­bed­ding ma­trix and pre­vi­ous-to­ken heads: how do they ac­tu­ally work?

  Here’s the plots you asked for for all heads! You can find them at:

  https://​​github.com/​​adamyedidia/​​resid_viewer/​​tree/​​main/​​experiments/​​pngs

  Haven’t looked too carefully yet but it looks like it makes little difference for most heads, but is important for L0H4 and L0H7.

• AdamYedidia 10 Aug 2023 5:16 UTC

  1 point

  0

  on: The po­si­tional em­bed­ding ma­trix and pre­vi­ous-to­ken heads: how do they ac­tu­ally work?

  The code to generate the figures can be found at https://​​github.com/​​adamyedidia/​​resid_viewer, in the experiments/​​ directory. If you want to get it running, you’ll need to do most of the setup described in the README, except for the last few steps (the TransformerLens step and before). The code in the experiments/​​ directory is unfortunately super messy, sorry!