Motahareh Pourrahimi

Launching Agentick 🤖🧠 A unified benchmark for training and evaluating general sequential decision-making agents. RL agents, LLMs, VLMs, hybrids, bots, and humans can all be evaluated on: same tasks. same seeds. same score. First result: no single agent dominates. 🧵

Beyond that, the bigger challenge is unsupervised discovery of those latent manifolds from activations + behavior: discovering the underlying representations that drive cognitive behavior without assuming labels for those latents in advance. (3/3)

What’s especially interesting is going beyond the final output variable, e.g., days of the week: asking what latent variables the model uses internally, how they are represented geometrically, and whether manifold-aware steering can causally test how they affect behavior. (2/3)

Really enjoyed Goodfire’s neural geometry posts. A few thoughts. The current examples mostly focus on explicit variables, e.g., days of the week, and show how their geometry in activation space mirrors the geometry of the model’s outputs/behavior. (1/3)

New paper 🚨 #ICLR26 Most world models predict the future from a past trajectory. But neuroscience suggests that such inference can instead be made from temporally independent experiences. We built the Episodic Spatial World Model (ESWM), a model that does exactly this:

So proud of this work co-led by @aspa1313 and @johanobandoc Fantastic collaboration with @AaronCourville & @pcastr Tldr; Constraining representations within isotropic Gaussian distribution massively improves performance during RL!

📎 Preprint: biorxiv.org/content/10.110…

Excited to give a talk at #Cosyne2026 about my PhD work! We show that RNNs trained on visual search converge on brain-like solutions, producing primate-like behavior and neural representations. Happy to chat if you're at Cosyne! 📅 March 15, 10:45–11 AM 📎 Preprint 👇

In this piece for @_TheTransmitter, I argue that ecological neuroscience should leverage generative video and interactive models to simulate the world from animals' perspectives. The technological building blocks are almost here - we just need to align them for this application.

If you are still at #NeurIPS2025 come check out my "Data on the Brain & Mind" Workshop poster, if you are into these tingz...

We’ll be at CCN next week with two posters, one contributed talk and one satellite talk! If you are interested in neural coding properties in primate hippocampus, modeling visual search, or cortical topography come and talk to us! See👇for details:

Announcing our Final Line-Up & Reduced Prices: Spring Summit on Fundamental Challenges for AI! bit.ly/springsummitai… We discuss core challenges & promising avenues in a day of high-profile keynotes & panels. Lunch & Reception included. Come & join us! April 15th, @ComputerHistory, Mountain View, CA. Keynotes Professor Jay McClelland (@Stanford / @GoogleDeepMind) Professor Bin Yu (@UCBerkeley / @ChanZuckerberg) Professor Diyi Yang (@Stanford; @Diyi_Yang) Professor Sergey Levine (@UCBerkeley; @svlevine) Panels New Landscapes for AI Research @cashewmake2 @AGamick @SGRodriques @joewhaley @fuelfive ++ New Professional Pathways @flxsosa @skylrwang @mayankagrawal ++ Challenges in BioMed @alexosh @JAAQOFFICIAL ++ Fundamental Technical Challenges @SuryaGanguli @Azaliamirh @ElizaKosoy ++ Speakers from @Google @Stanford @UCBerkeley @Harvard @MIT @Princeton @UniofOxford @Google @Meta @a16z @GoogleDeepMind @AIatMeta @stanfordnlp @StanfordAILab @berkeley_ai @StanfordHAI @StanfordEng @Berkeley_EECS @Convergent_FROs @FutureHouseSF @MITBrainandCog @SimonsInstitute @costplusdrugs @fiftyyears @netflix @czbiohub @neur_reps @BasisLongevity x.com/RMBattleday/st…

📢The UNIQUE Student Affairs Committee (SAC) is looking for motivated students to join the team! If you want to be part of the Quebec Neuro-AI community, the SAC is a golden opportunity for you! 🤖🧠 👉Fill out the form before August 17th : forms.gle/bajhxJLtU5YttM…

Had an amazing time wrapping up my last @BrainReach session this year! Teaching brilliant elementary kids about the brain was pure joy. One student wished for BrainReach 100 days a year—made my day! Excited to share this experience in my first-ever tweet!

Neural networks are making headlines everyday. The rapid progress is astonishing, sometimes exhausting… But are these improvements yielding more accurate neuronal models? @Yifei_Ren asked this question in a new preprint and the results were surprising biorxiv.org/content/10.110… 🧵

📢We are hiring ! We are looking for student to join the UNIQUE Newsletter team. ⌛️Fill out the application form until April 14th and join the team to explore the exciting research landscape of Neuro-AI happening in the labs all across UNIQUE! 👇docs.google.com/forms/d/e/1FAI…

Ever wondered about why and how cortical topography emerges in the cortex? Come by our poster at #cosyne2023 tomorrow evening (Saturday Mar 11; 8:30-11:30pm) at talk to the wonderful @great_caster about his work

@great_caster @PouyaBashivan Congrats!

I am thrilled to announce that our abstract on building a topographical neural network model of the visual cortex has been accepted to #cosyne2023! Huge thanks to my supervisor and mentor @PouyaBashivan for guiding me through this research. Excited to meet everyone there.