Twitxy Saved

Kai Sandbrink

@akaijsa

Computational cognitive neuroscience PhD Student @UniofOxford & @EPFL_en co-supervised by @summerfieldlab and @compneuro_epfl
Joined March 2023
202 Following 195 Followers 27 Posts

Latest posts 5

Kai Sandbrink @akaijsa
Excited to have been a part of this effort, and looking forward to what comes next!
James Whittington Verified account @jcrwhittington Aug 12
We’re excited to announce DiG-bench, a new benchmark for discovery!

Over the last few weeks we’ve been testing frontier AI models on our novel discovery games and seeing how they score. Each game is a text-based environment, so they probe discovery capabilities in the natural domain of language models, rather than requiring additional, potentially confounding, visual understanding.

TL;DR frontier models have improved a lot over the last few months. But they are still stumped by some surprisingly simple problems, even in their native text domain.

With

@cocosci_lab (@Princeton) @MITCoCoSci (@MIT) @SchmidhuberAI (@KAUST_News) @misovalko (@Inria) @tri_dao (@PrincetonCS) @RMBattleday @zebkDotCom @FraserGreenlee @akaijsa @ClareMaguire @TimMuller1 @kubicek_ales @physicscat0x7d @SukritSumant @thoughtchannel_

(1/5)
View quoted post on X
Kai Sandbrink @akaijsa
Excited that my paper on metacontrol in humans and neural networks with @summerfieldlab and @LHuntNeuro is out in PNAS! We examine the way that predictive representations of control enable behavioral adaptation across settings, and links to pathologies: pnas.org/doi/10.1073/pn…
www.pnas.org Understanding human metacontrol and its pathologies using deep neural networks | PNAS Error monitoring is crucial for inferring how controllable an environment is, and thus for estimating the value of control processes (metacontrol)....
3 replies 13 reposts 54 likes 5K views View on X
Kai Sandbrink @akaijsa
This was an amazingly fun collaboration that began at the AC Summer School @GatsbyUCL, and I can't wait to see where it takes us next! Check out our spotlight paper and poster at NeurIPS.

Grainy picture from presenting the first version of this work at @CosyneMeeting attached!
Thrilled to share our NeurIPS Spotlight paper with @japhba* @akaijsa* @SaxeLab @summerfieldlab @Ahummos*!

We study how task abstractions emerge in gated linear networks and show that these abstractions support cognitive flexibility.

openreview.net/pdf?id=AbTpJl7…
View quoted post on X
1 replies 1 reposts 21 likes 2.1K views View on X