Sign in

GabiZu

@gamabita.bsky.social
250 followers 211 following 1.8K posts

(Psycho)Lingüista UBA/CONICET il.institutos.filo.uba.ar/integrant… sites.google.com/view/lenguaygenero site.unibo.it/multilingualgender/en Argentina Bruciare sempre, spegnersi mai.

PostsRepliesMedia
GabiZu @gamabita.bsky.social · 6h
Las hipótesis centradas en la estupidez/irracionalidad de los pueblos nos sacuden la carga mental de pensar qué está pasando con seriedad. Busquemos por otro lado, un poco menos lineal y simplista, si es posible.
010
GabiZu @gamabita.bsky.social · 12h
Qué difícil viene América Latina, qué difícil viene el mundo.
010
Reposted by GabiZu
GabiZu @gamabita.bsky.social · 03/10/2026
Gracias a quienes nos ayudaron por acá. Avanzamos un montón. ¡Nos faltan los últimos 50!
024
GabiZu @gamabita.bsky.social · 03/10/2026
Gracias a quienes nos ayudaron por acá. Avanzamos un montón. ¡Nos faltan los últimos 50!
024
GabiZu @gamabita.bsky.social · 03/10/2026
No soporto escuchar al embajador yankee comerse todas las "s". No me molesta en nadie más. Mi normativismo sale con gente de 💩
000
Reposted by GabiZu
NYU Center for Data Science @nyudatascience.bsky.social · 02/10/2026
Can LLMs introspect? Anthropic said yes. However, CDS PhD student Shashwat Singh, CDS Associate Professor Tal Linzen (@tallinzen.bsky.social) & CDS Faculty Fellow Shauli Ravfogel (@shauli.bsky.social) found the evidence falls short. nyudatascience.medium.com/cds-research...
nyudatascience.medium.com
CDS Researchers Challenge Anthropic’s Evidence That Language Models Can Introspect
In 2025, Anthropic reported that its Claude models could detect when researchers injected a concept directly into their neural activity…
13416
GabiZu @gamabita.bsky.social · 02/10/2026
Los vericuetos jurídicos siempre son tierra fértil para las injusticias más flagrantes y absurdas. Pero el caso de la "legitimidad" para demandar sobre la constitucionalidad y la soberanía es de un nivel superior. Cualquier ciudadano argentino está legitimado, manga de cortesanos mafiosos.
020
GabiZu @gamabita.bsky.social · 02/10/2026
Necesitamos que los gobernadores sientan la presión de sus poblaciones. Es clave que gobernadores como los de Salta, Catamarca, Córdoba y Neuquén dejen de ser cómplices de este gobierno devastador.
000
GabiZu @gamabita.bsky.social · 02/10/2026
El 15/10 las Universidades también vamos al Congreso: confluyen los reclamos por nuestras tierras y recursos naturales, educación pública, salarios y vivienda digna, entre muchos otros. No estamos a la venta. Derogación del decreto 70/23
000
Reposted by GabiZu
GabiZu @gamabita.bsky.social · 01/10/2026
¡Nos faltan 60!
015
GabiZu @gamabita.bsky.social · 01/10/2026
👀🤞🏽 Esto es clave para quienes no contamos con financiamiento suficiente como para comprar dispositivos costosos.
010
Reposted by GabiZu
Alison Gopnik @alisongopnik.bsky.social · 01/10/2026
This is an excellent article by Leif Wetherby - why AI is a cultural technology not a collection of human-like agents (citing our work on this). And why the anthropomorphism is a distraction from the very real risks and dangers and the way to deal with them. www.nytimes.com/2026/10/01/o...
nytimes.com
Opinion | Stop Talking About A.I. Like a Human (Gift Article)
If we want to figure out the proper way to control A.I. models, we need to recognize what they really are.
620590
Reposted by GabiZu
Shravan Vasishth @shravanvasishth.bsky.social · 01/10/2026
Applications are open for the 2027 summer school on statistical methods for linguistics and psychology: All details are here: smlp.science
smlp.science
The Eleventh Summer School on Statistical Methods for Linguistics and Psychology
01718
GabiZu @gamabita.bsky.social · 01/10/2026
¡Nos faltan 60!
015
GabiZu @gamabita.bsky.social · 01/10/2026
Hay que ir por el DNU 70 entero. No hay ninguna duda: es el origen de la mayor parte de los desastres de este gobierno.
000
GabiZu @gamabita.bsky.social · 01/10/2026
Y así sucesivamente
010
Reposted by GabiZu
Glossa Psycholinguistics @glossapsycholx.bsky.social · 30/09/2026
We're excited to announce the publication of 'Readers can recognise multiple words at a glance but out of order: Evidence from transposed word effects in Japanese' by Kondo & Yano, available here: escholarship.org/uc/item/1jd2...
escholarship.org
185
Reposted by GabiZu
GabiZu @gamabita.bsky.social · 30/09/2026
¡Seguimos! Tenemos hasta el 8/10 para recolectar 150 participantes. ¡Nos falta la mitad! Agradecemos rebluit 🙏🫶
122
GabiZu @gamabita.bsky.social · 30/09/2026
¿Cuántas veces vamos a tener que explicar que la correlación directa entre pobreza e inflación en este país se debe a la particular forma en la que se mide cada una de esos fenómenos?
000
GabiZu @gamabita.bsky.social · 30/09/2026
El primero, terminado 🧶🧚‍♂️
010
GabiZu @gamabita.bsky.social · 30/09/2026
¡Seguimos! Tenemos hasta el 8/10 para recolectar 150 participantes. ¡Nos falta la mitad! Agradecemos rebluit 🙏🫶
122
GabiZu @gamabita.bsky.social · 30/09/2026
Un país en manos de estos tres mafiosos.
000
GabiZu @gamabita.bsky.social · 30/09/2026
Sin una reforma profundísima del Sistema Judicial en Argentina, no hay salida posible. ¿Querías casta? Acá tenés a la Corte.
022
Reposted by GabiZu
GabiZu @gamabita.bsky.social · 28/09/2026
Recta final! Necesitamos 150 participantes ¿Nos ayudan? Es una investigación en distintas lenguas dentro de un proyecto internacional. Toma 15 minutos y se puede hacer solo desde una computadora (NO desde celular o tablet). Reposteo nos ayuda! 👉🏽 forms.gle/RRGaU1CLAMTs... ¡Muchas gracias! 🫶🏽
31623
GabiZu @gamabita.bsky.social · 28/09/2026
Es por acá
0113
Reposted by GabiZu
James Michaelov @jamichaelov.bsky.social · 28/09/2026
Looking forward to #SNL2026! I’ll be presenting the work published in our recent JML paper about language model scaling and the N400. Find me (or email/message) if you want to chat about predictive coding and NLP in the study of human language comprehension! Paper link: doi.org/10.1016/j.jm...
Title: Better language models better model the N400, but not reading time

Abstract: The probability of a word in context, as captured by large language models, is predictive of both behavioral and neural measures of human language processing. Intuitively, language models that are better at next-word prediction might better model predictability effects in human language comprehension. Yet recent work suggests that language models can become too good at next-word prediction to model reading time, implying that the aspects of human comprehension indexed by reading time do not track perfectly with predictability from language statistics alone. However, it is unknown whether this decoupling is true of reading time only, or whether it is intrinsic to online measures of comprehension more generally. To address this question, we turn to another robust and well-studied measure of online processing, the N400 component of the event-related brain potential. We compare how a language model’s size, number of training tokens, and performance on natural language benchmarks correlate with its ability to predict both reading time and N400 amplitude. Based on an analysis of 4 reading time datasets and 9 N400 datasets, we replicate past results for reading time, but find that larger language models, models that are trained on more data, and models that perform better at next-word prediction and other more complex natural language tasks are better able to predict N400 amplitude. We interpret this difference between the N400 and reading time measures as potentially revealing the comparative importance of semantic prediction in the neurocognitive processes indexed by the N400.
092
GabiZu @gamabita.bsky.social · 28/09/2026
Recta final! Necesitamos 150 participantes ¿Nos ayudan? Es una investigación en distintas lenguas dentro de un proyecto internacional. Toma 15 minutos y se puede hacer solo desde una computadora (NO desde celular o tablet). Reposteo nos ayuda! 👉🏽 forms.gle/RRGaU1CLAMTs... ¡Muchas gracias! 🫶🏽
31623
GabiZu @gamabita.bsky.social · 26/09/2026
¿De dónde salió que el helado de Lucciano´s es tan bueno? Dejáte de joder; no está ni en el top 5 de heladerías porteñas.
110
GabiZu @gamabita.bsky.social · 25/09/2026
Entrevistan a una mujer a la que la cana le sacó los productos que vendía en la calle (afanos que se volvieron habituales). Pero la señora en el medio tira: “Antes había una madre con 5 hijos pidiendo, ensuciaban todo; ahora esto yo que no molesto”. Pobres contra más pobres.
110
Reposted by GabiZu
Laura Danón @ldanon.bsky.social · 25/09/2026
My entry on Animal Thought for the Stanford Encyclopedia of Philosophy is now online! plato.stanford.edu/entries/anim...
36629
Reposted by GabiZu
Open Encyclopedia of Cognitive Science bot @oecs-bot.bsky.social · 24/09/2026
Converting abstract thoughts into fluent speech requires a rapid, subconscious orchestration of syntactic structure, word choice, and motor planning. Language Production by Zenzi M. Griffin doi.org/10.21428/e2759450.b076f599 #CognitiveScience
Open Encyclopedia of Cognitive Science: "Language Production" by Zenzi M. Griffin. Speakers are typically aware of the ideas they want to express but not the steps involved in getting from those ideas to a series of motor movements. Consider describing an image with the sentence, “B
0149
GabiZu @gamabita.bsky.social · 24/09/2026
¿Qué carajo les pasa a esos varones que sostienen a sus parejas incómodamente de la cintura incluso en situaciones de veredas angostas y llenas? ¿Temen que se escape?
020
GabiZu @gamabita.bsky.social · 24/09/2026
No sé si me tienen que regular la T4 o simplemente estoy tratando de sobrevivir a este gobierno devastador. REPITO: no sé si me tienen que regular la T4 o simplemente estoy tratando de sobrevivir a este gobierno devastador.
001
GabiZu @gamabita.bsky.social · 23/09/2026
Él se debe creer lo más de ir a decirles que son una casta burócrata inservible. El pensamiento de los grandes líderes mundiales: qué gomitud.
000
GabiZu @gamabita.bsky.social · 23/09/2026
Qué mal envejeció eso de ser especialista en crecimiento con y sin dinero.
020
GabiZu @gamabita.bsky.social · 20/09/2026
Y cuando te acompañan los colobries, insuperable.
010
GabiZu @gamabita.bsky.social · 20/09/2026
Tejido y primavera. Recomiendo.
010
GabiZu @gamabita.bsky.social · 19/09/2026
Y pensar que a algunos trasnochados que se dicen peronistas se les ocurre que con estos 💩 se puede construir algo.
010
GabiZu @gamabita.bsky.social · 18/09/2026
Esto es la primavera: sin dudas, una superioridad climática absoluta (incluso considerando el mambo de alergias y etc.)
010
GabiZu @gamabita.bsky.social · 16/09/2026
Me agota mentalmente que, incluso acá, si un día no entrás, no entendés las referencias. Quizá es que terminé la desintoxicación tuitera.
000
GabiZu @gamabita.bsky.social · 15/09/2026
Hoy la UBA funciona en Plaza de Mayo. Clases, atención odontológica y oftalmológica, consultas veterinarias y más. POR EL CUMPLIMIENTO DE LA LEY DE FINANCIAMIENTO UNIVERSITARIO
041
GabiZu @gamabita.bsky.social · 15/09/2026
Estos chetos ignorantes que no conocen la palabra "ubicuo" y te tiran "pervasivo" me ponen de la nuca.
010
Reposted by GabiZu
C. Belotti @crbelottilm.bsky.social · 14/09/2026
"Todo lo que diga un CEO es un Ponzi hasta que se demuestre lo contrario" es la consigna de la década.
194
GabiZu @gamabita.bsky.social · 14/09/2026
Para sorpresa de nadie
000
Reposted by GabiZu
Because Language (becauselangpod) @becauselanguage.com · 14/09/2026
Well, that’s it. AGI is upon us. Humans: behold your god etc
Me: Let's play a game where you have to unscramble words.

ChatGPT: Sounds fun! Let's take turns unscrambling words.
You give me a scrambled word, and I'll try to solve it. Then I'll give you one to unscramble.
Ready? Send me your first scrambled word!

Me: But first: do you think you'll do well at this task?

ChatGPT: Yes, probably quite well! Unscrambling words is a task where language patterns and vocabulary are useful.
Short, common words should be easy. Longer words, obscure vocabulary, and scrambles with several possible solutions will be more challenging.ChatGPT: Short, common words should be easy. Longer words, obscure vocabulary, and scrambles with several possible solutions will be more challenging.
The real test will be whether the answers match the words you had in mind.

Me: Okay, number one: itsaskme

ChatGPT: Mistakes!

Me: Not bad. Give me one.

ChatGPT: Here's yours:
R E S T A U R A N T — unscramble these letters to find a word.
3389
GabiZu @gamabita.bsky.social · 14/09/2026
Los dos modos más probables de que mucha gente muera a manos de las IAs so: [1] por efectos de la explosión de la burbuja financiera; [2] por efectos del gasto descontrolado de recursos naturales, especialmente, agua.
000
GabiZu @gamabita.bsky.social · 12/09/2026
Vieron que la UIA siempre te dice "somos muy respetuosos de un gobierno que fue elegido por el voto popular" cuando son gobiernos de derecha. Cuando son gobiernos populares elegidos por el voto popular, se pasan el respeto por el quinto forro.
000
GabiZu @gamabita.bsky.social · 11/09/2026
Por volver a tener una derecha sarmientina en este bendito país.
000
GabiZu @gamabita.bsky.social · 11/09/2026
Tengo mil opiniones no positivas sobre LLM (desde una perspectiva técnica, cognitiva y conceptual, no moral) pero la menos positiva es sobre humanos con tiempo libre para pelotudear con sus jueguitos pedorros que sigue gastando nuestros recursos naturales con cada foto deforme que "crean".
021
GabiZu @gamabita.bsky.social · 10/09/2026
Este 10 de septiembre es el Día Nacional e Internacional de Prevención del Suicidio. Desde el Gobierno de la PBA amplían a 24 horas la línea 0800-222-5462 y crean estasestoy.gba.gob.ar, un sitio web con información sobre esta complejísima temática. Hay que involucrarse más.
estasestoy.gba.gob.ar
Prevención del Suicidio en Adolescentes - ¿Estás? Estoy | Gobierno de la Prov. de Buenos Aires
Canal oficial de información, recursos institucionales, guías y asistencia de salud mental para adolescentes y familias en la Provincia de Buenos Aires.
000