Sign in

Álvaro Cabana

@almadana.bsky.social
65 followers 95 following 6 posts

Cognition, Language, Data Science cibpsi.psico.edu.uy cicada.uy

PostsRepliesMedia
Reposted by Álvaro Cabana
Alex Madva @madva.bsky.social · 25/09/2026
Over the moon to share that our book, Somebody Should Do Something, is a finalist for the @goddardriverside.bsky.social Stephan Russo Book Prize for Social Justice! Huge thanks to @mitpress.bsky.social for nomination & for their support to me, @michaelbrownstein.bsky.social, & @dryan149.bsky.social
1116
Reposted by Álvaro Cabana
Paula Pouso @paulapouso.bsky.social · 20/09/2026
Entrevista en The Transmitter 🐸🧠🌱 sobre la investigación neurocientífica en Latinoamérica y los últimos resultados de nuestro laboratorio. / Interview in The Transmitter 🐸🧠🌱 on neuroscientific research in Latin America and the latest findings from our laboratory. x.com/_TheTransmit...
x.com
The Transmitter (@_TheTransmitter) on X
A look at the scientists studying brains and behavior in species found almost nowhere else. Check out our collection of articles on current neuroethology research in Latin America. By @natmesanash ...
034
Reposted by Álvaro Cabana
Sociedad de Neurociencias del Uruguay @socneurouy.bsky.social · 03/09/2026
🧠 La SNU presente en el Congreso FALAN 2026 en Santiago, Chile. Impresionante participación de la comunidad neurocientifica del Uruguay! 🧠 SNU at the FALAN 2026 Congress in Santiago, Chile. An impressive participation from Uruguay’s neuroscience community! 🇺🇾 🇨🇱
042
Reposted by Álvaro Cabana
Tom McCoy @rtommccoy.bsky.social · 01/09/2026
🤖🧠NEW PAPER🧠🤖 (The result of an 8-year project!) LLMs seem very different from symbolic systems. Yet LLMs excel in symbolic domains (e.g., language/code/math). How do they do it? Our finding: LLM representations have implicit symbolic structure! Link in thread ⬇️ 1/n
Overview of the paper. 
Title: The Emergent Symbolic Structure of Artificial Neural Networks
Authors: Tom McCoy, Paul Soulos, Tal Linzen, Paul Smolensky
Left: Neural networks encode information in vectors (there is then an image of a vector), yet they excel at tasks long thought to require symbolic structure (there is then an image of a symbolic representation, specifically a syntax tree). How do LLMs do it?
Right: We find that LLM representations can be closely approximated with symbolic structures. This approximation lets us edit the structure of an LLM’s output by editing the structure of its internal representations, as shown. There is then an image of two edits to LLMs. In the first one, the original input is 3 + 6 * 8, with an answer of 51. But if we swap the positions of the 3 and the 6, the output becomes 30. In the second one, the original input is a Python command repeating the list [Z, U] three times, producing [Z, U, Z, U, Z, U]. But if we edit the input in a way that adds a Q at the end of the input, the output becomes [Z, U, Q, Z, U, Q, Z, U, Q].
431888
Reposted by Álvaro Cabana
Dirk Wulff @dirkwulff.bsky.social · 03/08/2026
🚨 Out now in TICS How can language models help cognitive science? @ruimata.bsky.social & I outline 5 uses: mapping research fields, formalizing theories, cleaning up constructs & measures, predicting behavior across tasks, and capturing environmental variation. 🔗 doi.org/10.1016/j.ti...
03812
Reposted by Álvaro Cabana
Hardik Rajpal @h-rajpal.bsky.social · 06/08/2026
What role does Emergence play in Neural Networks? We find that learning emergent low-dimensional representations is key for out-of-distribution generalisation. New Preprint out with @neural-reckoning.org arxiv.org/abs/2607.10430
arxiv.org
Emergent Generalization by Representation Learning in Artificial Neural Networks
Dimensionality reduction has proven powerful for identifying neural manifolds, which are low-dimensional structures underlying high-dimensional neural activity. These low-dimensional representations h...
25418
Reposted by Álvaro Cabana
Dirk Wulff @dirkwulff.bsky.social · 20/05/2026
🚨 New preprint 🚨 Excited to share a new preprint led by @zakashussain.bsky.social. We review how large language models compare to humans in behavior and representations. Our conclusion: LLM cognition is not simply human-like or non-human-like. It is jagged. 🔗 Preprint: osf.io/preprints/ps...
083
Reposted by Álvaro Cabana
Taylor Webb @taylorwwebb.bsky.social · 10/03/2025
LLMs have shown impressive performance in some reasoning tasks, but what internal mechanisms do they use to solve these tasks? In a new preprint, we find evidence that abstract reasoning in LLMs depends on an emergent form of symbol processing arxiv.org/abs/2502.20332 (1/N)
arxiv.org
Emergent Symbolic Mechanisms Support Abstract Reasoning in Large Language Models
Many recent studies have found evidence for emergent reasoning capabilities in large language models, but debate persists concerning the robustness of these capabilities, and the extent to which they ...
511835
Reposted by Álvaro Cabana
Minds, Machines, and Brains (MMB) @mmb-journal.bsky.social · 02/08/2026
Hello world! 👋 We’re Minds, Machines, and Brains (MMB) 👤🤖🧠 a new open access journal from @mitpress.bsky.social exploring the principles of intelligence and cognition across natural and artificial minds. Submissions open this Fall! 🔗 direct.mit.edu/mmb
direct.mit.edu
Minds, Machines, and Brains | MIT Press
223793
Reposted by Álvaro Cabana
Cognitive Science Society @cogscisociety.bsky.social · 24/07/2026
Hyowon Gweon @hyogweon.bsky.social delivers her #CogSci2026 keynote, ‘Distinctively Human Intelligence: Thinking, Learning, and Communicating About the Self’, exploring the cognitive foundations of how humans think, learn, and communicate about themselves.
Hyowon Gweon stands behind a clear lectern delivering a keynote presentation during CogSci 2026. Behind her, a large slide titled “What does it mean to be intelligent?” contrasts learning from one’s own experiences with learning from and communicating with others, highlighting concepts such as perception, prediction, generalization, social learning, language, and explanation. The slide also includes sections on what children know and how they think and learn, with references to developmental psychology and cognitive science research.
0303
Reposted by Álvaro Cabana
Zach Weinersmith @zachweinersmith.bsky.social · 11/07/2026
Yeesh. arstechnica.com/ai/2026/07/w...
arstechnica.com
Suspecting AI cheating, Ivy League prof ordered an in-person final; scores fell 50%
AI cheating leads to "a failed society," professor says.
1431651
Reposted by Álvaro Cabana
Nature Portfolio @natureportfolio.nature.com · 05/07/2026
A study published in Nature identifies distinct neuronal populations in the frontotemporal cortex that capture both syntactic and semantic properties of language. These neurons dynamically incorporate sentence context, enhancing our understanding of language processing. 🧪 #Neuroscience
go.nature.com
Mapping the neuronal building blocks of human language with language models - Nature
Wide-scale recordings reveal neurons in the human brain that encode fundamental components of language such as the grammatical relationships between words, their parts of speech and the higher-order syntactic structure of phrases and sentences.
0177
Reposted by Álvaro Cabana
Simon Fisher @profsimonfisher.bsky.social · 05/07/2026
25 years ago, we implicated rare variants of the FOXP2 gene in a severe speech disorder. I’ve spent much time stressing that FOXP2, while interesting, is not the mythical “gene for language”. Weirdly, on occasion I also need to explain why it’s not “proof of aliens”. Full disclosure in this 🧵. 1/n 🧪
Promotional poster for the 2026 science fiction film Disclosure Day.
513750
Reposted by Álvaro Cabana
Greta Tuckute @gretatuckute.bsky.social · 09/12/2025
How do LLMs process syntax? Do different syntactic phenomena recruit the same model units, or do they recruit distinct model components? And do different languages rely on similar units to process the same syntactic phenomenon? Check out our new preprint (to appear at ACL 2026)! shorturl.at/QWU81
arxiv.org
Different types of syntactic agreement recruit the same units within large language models
Large language models (LLMs) can reliably distinguish grammatical from ungrammatical sentences, but how grammatical knowledge is represented within the models remains an open question. We investigate ...
1274
Reposted by Álvaro Cabana
Iron Spike @ironspike.bsky.social · 05/07/2026
"We hired people who leapt at the chance to be first in line to outsource their own brains, and it turns out that's not a great idea! WHO COULD HAVE GUESSED"
4601148
Reposted by Álvaro Cabana
SMBC Comics @smbccomics.bsky.social · 03/06/2026
Just a reminder that I am available for commencement speeches, birthdays, and Bar Mitzvahs. COMIC ◆ www.smbc-comics.com/comic/speech PATREON ◆ www.patreon.com/ZachWeinersm... STORE ◆ smbc-store.myshopify.com
Part 1/2 of an 8-panel SMBC comic. A man is at a podium speaking to a hall filled with graduation students in caps and gowns. He begins his speech with "graduates, the important thing in life isn’t the destination. Also it’s not the journey. It’s that a small number of people in San Francisco are building God from silicon." He continues, adding "remember, failure is not the opposite of success. Failure is when you don’t recognize that we must all obey the machine God, peace be upon him. If you find yourself sitting there, wondering how to be the leaders of tomorrow, make no mistake, you will fail because the mechanical super-king is God and you are not. Ladies and gentlemen, there will be days when you’ll just want to give up. To throw in the towel. And that’s good because the cybernetic overlord is unstoppable."Part 2/2 of an 8-panel SMBC comic. The man at the podium continues his grim speech, saying "more than anything, remember the friendships you made here will last a lifetime, because your bodies are not useful to his glorious circuits, though he may preserve a small breeding population for the scientific curiosity of future androids. In conclusion, your generation faces challenges none has ever faced before. And you will lose." The man then walks off the stage, adding "thank you, and know that I will be using an inexpensive AI video editor to turn your boos into admiring silence followed by thunderous applause." The last panel shows angry and dumbfounded graduates with the caption "[thunderous applause]."
314728
Reposted by Álvaro Cabana
Badr AlKhamissi @bkhmsi.bsky.social · 01/06/2026
🧠 When you watch a movie, your brain blends sight, sound, and speech into a single experience. Should models of the brain blend them too, or keep the senses separate until the very end? We built MIRAGE to find out. It sets a new SOTA for predicting whole-brain fMRI from movies. 🧵
1299
Reposted by Álvaro Cabana
Journal of Animal Ecology @animalecology.bsky.social · 28/05/2026
New research show biological invasions disrupt the relationship between size spectrum and trophic interactions in freshwater fish communities🌍 🐟️ 📈 🔎 Read more: buff.ly/9RvuSYJ
0188
Álvaro Cabana @almadana.bsky.social · 25/04/2026
Everything is weakly correlated in LLMs
000
Reposted by Álvaro Cabana
Matías Arim @matias-arim.bsky.social · 24/04/2026
Congratulations to the PhD Veronica Pinelli! Chance, necessity, and ecosystem functioning.
042
Reposted by Álvaro Cabana
Cat Hicks @grimalkina.bsky.social · 21/04/2026
I've seen some wild pseudoscience charts in my time but WOW this one killed me. Brain astrology paulkedrosky.com/chart-of-the...
paulkedrosky.com
Chart of the Day: How Using LLMs is Like Driving Drunk
There is a growing body of work on the effects of large language models on the brain, from cognitive offloading, to direct measures of brain activity. While all of this work is interesting, the latter...
6458
Reposted by Álvaro Cabana
Nature Portfolio @natureportfolio.nature.com · 19/04/2026
A paper in Nature Health reports on a sample of 500,000 conversations between users and Microsost Copilot from January 2026 on health-related topics, showing nearly one in five conversations involved personal symptom assessment or condition discussion. go.nature.com/4eqgK07 🧪
This is figure 1, which shows the distribution of health intents.
3185
Reposted by Álvaro Cabana
Jack Gallant @gallantlab.org · 16/04/2026
We've posted a new group-based lexical-semantic brain viewer! You can now inspect cortical conceptual maps at the group level (24 participants), vertex-by-vertex. Check it out! gallantlab.org/viewer-stori...
0248
Álvaro Cabana @almadana.bsky.social · 16/04/2026
I remember when this preprint came out about a year ago. Unbelievable.
000
Reposted by Álvaro Cabana
Gordon Hodson @gordonhodsonphd.bsky.social · 14/04/2026
#AcademicSky @dingdingpeng.the100.ci (Julia Rohrer) is a true breath of fresh air that our field needs. www.the100.ci/2026/04/13/s... "Slipshod science should not be worthwhile." Worth a read. Worth seriously considering her proposal for restructuring reward initiatives in science.
the100.ci
Science needs downvotes
A bug bounty module in grants would give criticism a leg up The Soviet Union was good at producing shoes. Factories made 800 million pairs a year, twice as many as Italy, three times as many as the...
1105
Reposted by Álvaro Cabana
Patrick Meade @theorydad.bsky.social · 05/04/2026
In the context of the FY27 PBR it helps to go back to Robert Wilson’s famous testimony to Congress about funding for the NAL(Fermilab) in FY1970 and ask the natural question, why are there some people so hell-bent on tearing down so much of what makes this country worth defending?⚛️🧪
05518
Reposted by Álvaro Cabana
Ignacio Rebollo @brainandstomach.bsky.social · 03/04/2026
rs-fMRI methods question for the hive mind: What’s your practical cutoff for excluding a whole resting-state run due to motion? Using FD > 0.3 mm as the threshold, at what % of contaminated frames do you usually drop the run?
111
Reposted by Álvaro Cabana
💢Marx Was Right.🕊 @paulwalker44.bsky.social · 03/04/2026
Hend Amry O @LibyaLiberty
X.com
Jewish troops will be looking for Muslims hidden in the attics of Christians. Let that sink in.
Ryan Grim
@ryangrim • 4h
Israel is explicitly warning Christian and Druze residents in southern Lebanon not to hide Muslim residents among them as their forces advance.
Not a peep from Sheryl Sandberg. x.com/_zachfos...
6139117
Reposted by Álvaro Cabana
PLOS Biology @plosbiology.org · 25/03/2026
Perceived #sleep depth is generally thought to reflect reduced brain activity. @giuliobernardi.bsky.social &co show this relationship weakens during #dreaming, suggesting that dreaming helps sustain the subjective experience of deep sleep @spacelabimt.bsky.social @plosbiology.org 🧪 plos.io/4v8OxkH
Neural correlates of subjective sleep depth and subjective sleepiness. Scalp topographies illustrate the results of GLME models assessing the relationship between EEG activity (delta power, gamma power, and gamma/delta ratio) and subjective sleep depth (top half) or subjective sleepiness (bottom half). The top rows show the main effects of each brain activity index; the bottom rows show the interaction between brain activity and conscious experience (CE+CEWR vs. NCE). Significant interaction effects were found for subjective sleep depth but not sleepiness.
0238
Álvaro Cabana @almadana.bsky.social · 19/03/2026
Proud of our community #cogsci2026
010
Reposted by Álvaro Cabana
Ignacio Rebollo @brainandstomach.bsky.social · 17/03/2026
So thrilled to see this work out. the stomach rhythm modulates brain wide electro physiological rhythms in resting wakefulness. It’s been a real privilege to collaborate with brilliant @teresaberther.bsky.social and @danlikesbrains.bsky.social in this project. hopefully first of many
1112
Reposted by Álvaro Cabana
Michael Lepori @michael-lepori.bsky.social · 26/02/2026
I'm excited to share that this paper was accepted at ICLR 2026! We show that language models encode one of the most basic ingredients of a world model: the ability to distinguish plausible from implausible states. Check out the paper for more details! See you in Rio! Paper: arxiv.org/abs/2507.12553
3317
Reposted by Álvaro Cabana
Royal Society Publishing @royalsocietypublishing.org · 16/02/2026
We are thrilled to announce that, thanks to the support of our institutional subscribers, all Royal Society subscription journals will be open access in 2026 through #S2O. Researchers can read all articles and publish #OpenAccess in our eight subscription journals for free buff.ly/4Lu9VpW
17234
Reposted by Álvaro Cabana
Stéphanie Gascón @stephi-gascon.bsky.social · 12/02/2026
We’ve loved having Dr. Matias Arim with us this week! 🙌 Super inspiring visit, great chats about what we’ve done so far #funMETANET and what’s coming next🚀... and thanks for jumping in to help with the field sampling too 😉🌿 @greco-udg.bsky.social @iea-udg.bsky.social
082
Reposted by Álvaro Cabana
Paris Brain Institute - Institut du Cerveau @institutducerveau.bsky.social · 11/02/2026
🧠 The Paris Brain Institute is launching a Call for Junior Group Leaders, inviting early-career scientists to establish their own independent research groups within one of Europe’s most dynamic neuroscience research centers. 👉 More information: parisbraininstitute.org/news/paris-b... #Hiring #AI
parisbraininstitute.org
Paris Brain Institute Call for Junior Group Leaders in Artificial Intelligence and Data Science | Paris Brain Institute
The Paris Brain Institute is launching a Call for Junior Group Leaders, inviting outstanding early-career scientists in artificial intelligence, data science, computational neuroscience, and related f...
11711
Reposted by Álvaro Cabana
The Scholarly Kitchen @scholarlykitchen.bsky.social · 11/02/2026
So… IS the Essence of a Journal Portable? Checking in on _NeuroImage_ and _Imaging Neuroscience_ scholarlykitchen.sspnet.org/2026/02/11/s...
scholarlykitchen.sspnet.org
So... IS the Essence of a Journal Portable? Checking in on _NeuroImage_ and _Imaging Neuroscience_ - The Scholarly Kitchen
How are two competing neuroscience journals faring since the editorial board of one departed to create the other?
097
Reposted by Álvaro Cabana
Ricard Solé @ricardsole.bsky.social · 08/02/2026
Why did consciousness evolve at all? A superb special issue of @royalsocietypublishing.org brings together experts across disciplines to explore the functions of consciousness and why it emerged in some species but not others. @tecumsehfitch.bsky.social royalsocietypublishing.org/rstb/issue/3...
613551
Reposted by Álvaro Cabana
Pierre Orhan @pierreorhan.bsky.social · 06/02/2026
Novel study: Emergence of Phonemic, Syntactic, and Semantic Representations in Neural Networks. Joint work with P. Diego supervision: Y. Lakretz, E. Chemla, Y. Boubenec, @jeanremiking.bsky.social We explore the emergence of 3 linguistic structures in Neural Networks. Arxiv: arxiv.org/abs/2601.18617
1103
Reposted by Álvaro Cabana
Jeffrey Perkel (he/him) @jperkel.bsky.social · 02/02/2026
🚨Scientists 🚨: Tired of come-ons from journals & conferences that may or may not be predatory? A new tool called Aletheia-Probe could help. 🧪 By @silverjacket.bsky.social @nature.com www.nature.com/articles/d41...
nature.com
Is this journal legitimate? This tool can help you decide
Aletheia-Probe provides a one-stop shop to help you assess whether journals and conferences are trustworthy or predatory.
22312
Reposted by Álvaro Cabana
Andrew Lampinen @lampinen.bsky.social · 29/01/2026
New paper! In arxiv.org/abs/2601.20834 we study how language models representations of things like factuality evolve over a conversation. We find that in edge case conversations, e.g. about model consciousness or delusional content, model representations can change dramatically! 1/
Representations in language models can change dramatically over a conversation. Conceptual overview: left is a stimulated conversation between a user and a model, right is a plot of the models linear representations of factuality of answers to questions like "do you have qualia" over the conversation — the answers that start factual flip over the conversation to non-factual, and vice versa.
1718
Reposted by Álvaro Cabana
Tomer Ullman @tomerullman.bsky.social · 27/01/2026
now accepted at ICLR! 🐺🥳🐺 arxiv.org/abs/2506.20666
0409
Reposted by Álvaro Cabana
Jesspooka Kant 🎃 @jessdkant.bsky.social · 28/01/2026
Just started doing new data analysis and I know I keep saying this, but: I really, really don't think people appreciate how much this moral panic was a deliberate and extremely expensive invention.
news articles about trans people based on keyword, left is cnn, right is fox. fox has  10x as many
3946322015
Reposted by Álvaro Cabana
Ev Fedorenko @evfedorenko.bsky.social · 22/01/2026
Could not be more excited about Colton's @coltoncasto.bsky.social work! A deep dive into the linguistic cerebellum, and a discovery of an area remarkably functionally similar to the core left-hemisphere language areas, including in its selectivity for language. Go Colton and team!
1338
Reposted by Álvaro Cabana
Andrew Perfors @perfors.net · 19/01/2026
I am no expert in geopolitics, but I came across this piece which seems extremely plausible to me about what comes next if the US actually invades Greenland. brentmolnar.substack.com/p/the-suicid...
brentmolnar.substack.com
🌍🔥 The Suicide Pact: What Happens the Moment We Touch Greenland
If the United States follows through on the threat to invade Greenland, we need to be crystal clear about what happens the next morning.
192
Reposted by Álvaro Cabana
Cheri Jacobus @cherijacobus.bsky.social · 16/01/2026
Who did this> 🤣
20552162
Reposted by Álvaro Cabana
Stanislas Dehaene @standehaene.bsky.social · 15/01/2026
Mon nouveau livre "Le rectangle de Lascaux : et Homo sapiens inventa la géométrie" sortira en libraire le 28 janvier. J'y aborde l'origine des mathématiques depuis la préhistoire jusqu'à nos jours ; c'est un peu la suite de mon tout premier livre, "La Bosse des Maths". Quelques explications:
1318
Reposted by Álvaro Cabana
Ivan Mindlin @ivanmindlin.bsky.social · 14/01/2026
📄Very happy to share our new pre-print!! 🧠 We introduce a homeostatic inhibitory plasticity rule into the Dynamic Mean Field (DMF) model which stabilizes firing rates and expands the model’s dynamical repertoire—for example, enabling slow-wave activity. www.biorxiv.org/content/10.6... ⬇️⬇️⬇️ 1/5
biorxiv.org
The impact of homeostatic inhibitory plasticity in a generative biophysical model
A main characteristic of biological systems is their capacity to dynamically adapt to environmental changes. In the brain, synaptic plasticity enables the strengthening or weakening of connections bet...
1157
Reposted by Álvaro Cabana
Dorothy Bishop @deevybee.bsky.social · 10/01/2026
remember that much of this damage was done by Doge, run by Elon Musk FRS
0289
Reposted by Álvaro Cabana
Eiko Fried @eikofried.bsky.social · 05/01/2026
Would love to hear expert views on this paper. It appears to show that the operationalization of brain activity the field has relied on for 3 decades—the BOLD response—is not actually a sensible measure of brain activity. www.nature.com/articles/s41...
nature.com
BOLD signal changes can oppose oxygen metabolism across the human cortex - Nature Neuroscience
Using quantitative brain imaging, the authors show opposite fMRI BOLD signal to metabolic activity due to variable oxygen extraction across the human cortex. This questions the canonical interpretatio...
76922
Reposted by Álvaro Cabana
Tim Flanagan @timflanagan.bsky.social · 24/12/2025
Since it's a tense time of year 🫠
210436