Botos Csabi

Botos Csabi

@botoscsabi.bsky.social

Fresh DPhil from @oxfordtvg.bsky.social Working on low-latency/low-resource AI Neuroscience wannabe https://botcs.github.io/

51 followers52 following11 posts

Pinned

Botos Csabi
@botoscsabi.bsky.social

[Hugging Face release] We’re releasing the dataset behind Reason to Play! I’m incredibly proud of our team for reaching this milestone.

View full thread
Botos Csabi
@botoscsabi.bsky.social

... than the most sophisticated cognitive theory based models. There's no incentive for this stuff to happen. I think there may be some abstract universal circuitry that all intelligent systems have to independently rediscover.

View full thread
Botos Csabi
@botoscsabi.bsky.social

I have a really hard time wrapping my head around how and why frontier models would ever show human-brain alignment on games (unlike perception tasks, see Brain Score leaderboard)

Botos Csabi
@botoscsabi.bsky.social

they were trained on coding, math and all the domains, where the high validation score often meets extreme scrutiny due to benchmaxxers, and I just can't imagine that the labs went out of their way to train their models so that they end up predicting brain signals better ...

Botos Csabi
@botoscsabi.bsky.social

Jeee 🐦‍⬛ I am very proud of our joint effort with @sreejan.bsky.social on the project "Reason to Play" LRMs show human-like rule discovery, and their hidden states predict human brain activity during gameplay 10x better than previous methods Interactive demo + paper: botcs.github.io/reason-to-pl...

botcs.github.ioReason to Play - Behavioral and Brain Alignment32 fMRI-scanned humans and 8 frontier open weight LLMs play ARC-AGI like games with no rules given. The reasoning models match the human learning trajectories and their hidden states predict human bra...

You are up to date.