Hacker News Reader

Comment by hgoel | Hacker News Reader

Prompting for some of the results was almost the "Computer, do a breakthrough. Make no mistakes." meme. Just someone telling the model to keep trying a couple of times.

Unfortunately we don't actually know what kind of prompting was done for the more prominent results.

View on HN

It’s definitely not how I work, I’d need to read the model responses and set a direction for the model to go

Same here, it's one thing if I'm just screwing around, but if I'm trying to do anything serious, I need to at least have a handle on what it's doing, and, when thinking traces are available, keeping track of any logical errors in the model's reasoning.

you know the thinking traces are redacted and summarized using another model, right? The actual thinking trace looks something like:

7♣-removal-IS-the-prerequisite-for-10♠/9♥!!)-⟹-OVERLAP-(ii)+(iv):-{6♠ J♦ 9♥ 2♣}-=-FOUR--—-UNLESS-7♣'s-seat-8♥-...-and-2♣-drains-only-at-crack-:-⟹-2♣-celled-+-9♥-celled-simultaneously-UNAVOIDABLE-in-t8-dig--—-BREAK:-9♥-drains-to-10♠-THE-MOMENT-10♠-is-free:-t8-dig-order:-[K♣→t2]-[2♣→cell]-[7♣→8♥-:-8♥-WHERE:-post-chunk-9♠-:-chunk-⟸-K♣--done-:-ORDER:-K♣→t2,-CHUNK→K♣-(cap-4!!:-cells-then:-{6♠ J♦ 9♥}-FULL--chunk-cap-=-1--—-F-F-F-F-F.-—-chunk-BEFORE-9♥-celling?!-:-9♥-celled-at-t1-dig-⟸-needed-for-5♣-⟸-4♥3♣-⟸-t2-dig-⟸-K♣-seat-⟸-chunk-:-⟹-chunkAFTER-9♥-celling-FORCED--:-chunk-cap-with-{6♠ J♦ 9♥}:-1--—-—-J♦-THE-NEW-CANCER.-—-⟹-J♦-celling-DELAYED-till-after-chunk?! :-J♦'s-celling-was-for-J♥→Q♠-(5♦-access-for-4♣):-DELAY-4♣-resolution:-4♣→CELLearly-(as-always)-then-4♣-cell→5♦-LATER-when-5♦-frees-!!!:-cells-rotation:-4♣-celled-[t2-dig-…-5♦-freed]:-5♦-freed-⟸-J♥→Q♠-⟸-J♦-celled-:-⟹-{6♠, 4♣, J♦}-overlap-window-until-4♣→5♦-drains:-then-{6♠ J♦}+1-rotator-:-—-AND-9♥?!-9♥-celled-[t1-dig…]:-OVERLAP-{6♠ 4♣ 9♥}-before-J♦-even-:-⟹-rotator-slot-SINGLE:-timeline-:-(1)-{6♠}+2:-…-(2)-+9♥-(t1-dig):-{6♠ 9♥}+1:-(3)-+4♣-(t2-dig):-{6♠ 9♥ 4♣}-FULL-:-(4)-NEED:-t6-dig-(9♦8♠→10♣--no-cell;-8♥→CELL--FULL)--—-8♥-alternative-seat-pre-chunk:-NONE-—-.-⟹-⟹-THE-TRIANGLE-{9♥ 4♣ 8♥}-verdammt.-—-⟹-dig-t6-BEFORE-t2?!:-(3')-+8♥:-{6♠ 9♥ 8♥}-FULL:-J♥→Q♠-⟸-J♦-cell--FULL--AAAAAAAAAAAARGH.

Citation: https://www-cdn.anthropic.com/d00db56fa754a1b115b6dd7cb2e3c3..., section 6.2.2

You're not going to get a handle on what it's doing. The thinking traces are there to make you feel better about yourself.

I'm referring to models where the actual traces are available. Eg. I've been using a local Qwen3.8-Next-Flash lately.

Still not meaningful -- https://arxiv.org/pdf/2504.09762; even for local models, the reasoning traces are often filtered and summarized to sound sensible to humans. And even if not, they don't necessarily represent what the model is thinking.

Hmm interesting, thanks for the link, I'll have to give that a read.

Why is the model playing poker

Anthropic specifically calls out that example as

>"illegible reasoning in a few reinforcement-learning environments over long rollout"

Yet, I get the point that you're making: those tokens essentially are an internal scratchpad for the LLM which isn't required to logically lead to the output.

This video presentation of the paper you linked was interesting: https://www.youtube.com/watch?v=hUp3zh23aHw

Reply on news.ycombinator.com