Sign in

Iban Dlank (b↔︎d)

@ibandlank.bsky.social
732 followers 265 following 60 posts

Cognitive (neuro)scientist studying language @ UCLA Psych (BlankLangLab.com). Views are my own. He/Him. RT ≠ endorsement. 🏳️‍🌈

PostsRepliesMedia
Reposted by Iban Dlank (b↔︎d)
Dr Anna Leshinskaya @annaleshinskaya.bsky.social · 02/10/2026
I am recruiting graduate students for UCI's PhD in Cognitive Sciences - www.cogsci.uci.edu/graduate/pro... The lab's focus is on how relational structure and combinatorial reasoning is implemented in human brains and large language models--and if they are similar.
cogsci.uci.edu
GRADUATE PROGRAM
13430
Reposted by Iban Dlank (b↔︎d)
Rachel Ryskin @ryskin.bsky.social · 02/10/2026
These are great opportunities that can lead to tenure-track positions in the UC system. Happy to support a PPFP application if there’s a good research fit!
089
Reposted by Iban Dlank (b↔︎d)
Kanishka Misra @kanishka.bsky.social · 25/09/2026
Come be my colleague and teach meaning at UT Austin Linguistics!
078
Iban Dlank (b↔︎d) @ibandlank.bsky.social · 24/09/2026
My condition was that I’m not teaching it alone before tenure, I’ll only co-teach :)
001
Iban Dlank (b↔︎d) @ibandlank.bsky.social · 24/09/2026
I don’t use automatic AI detectors :) the distribution of emails and statements I get is currently very bimodal. The question is not who uses it as a drafting aid, but who uses it to offload thinking.
010
Iban Dlank (b↔︎d) @ibandlank.bsky.social · 24/09/2026
Tomer’s advice (including in the original post) strongly resonates with me! Also, re: gen AI—if you can’t be bothered to write your own statements, it indicates you won’t do your own research in grad school, and I think it’s a waste of my time to mentor someone like that.
150
Iban Dlank (b↔︎d) @ibandlank.bsky.social · 24/09/2026
The number of times I tried to convince our department to do this…
130
Iban Dlank (b↔︎d) @ibandlank.bsky.social · 21/09/2026
Receiving what I'm 99% sure are AI-generated emails from students in my lab, and AI-generated replies to comments on manuscripts. All due respect to increasing productivity, I'd much rather mentor humans. This is SO insulting.
040
Reposted by Iban Dlank (b↔︎d)
Arthur Prat @arthurpr4t.bsky.social · 09/09/2026
We derived six laws of psychophysics from a single efficient-coding equation. The laws include Weber's law; scaling laws in visual working memory (which had been noted before, but with unclear theoretical justification); Wei & Stocker's law of human perception; arthurprat.com/pdfs/Prat-Ca...
arthurprat.com
49628
Reposted by Iban Dlank (b↔︎d)
Kanishka Misra @kanishka.bsky.social · 08/09/2026
I am looking for two emergency reviewers on standby for ARR August. Topics include theory of mind interp in LMs and grammar induction! Please DM/email
003
Iban Dlank (b↔︎d) @ibandlank.bsky.social · 08/09/2026
Apparently I've been doing it wrong.
Text saying "Assistant Professor roles with over $550K/yr salary" from Linkedin
162
Reposted by Iban Dlank (b↔︎d)
Russ Poldrack @russpoldrack.org · 03/09/2026
We are searching for a tenure-track Assistant Professor position, with a research focus on developmental and/or social psychology. Please apply or forward to anyone who might be interested! facultypositions.stanford.edu/jobs/assista...
facultypositions.stanford.edu
Assistant Professor, Psychology - Stanford University, California, United States
The Department of Psychology at Stanford University is seeking applicants for a tenure-track Assistant Professor position, with a research focus on developmental and/or social psychology. In addition ...
08678
Reposted by Iban Dlank (b↔︎d)
aaron bornstein @aaronbornstein.bsky.social · 03/09/2026
UC Irvine Cognitive Sciences is hiring! Focus on social cognition, broadly construed. Come be my colleague :) I am not on the search committee but happy to talk about the campus, environment, careers, etc any time. recruit.ap.uci.edu/JPF10411
recruit.ap.uci.edu
Assistant Professor - Cognitive Sciences
University of California, Irvine is hiring. Apply now!
15863
Reposted by Iban Dlank (b↔︎d)
Liina Pylkkänen @liinapy.bsky.social · 16/08/2026
✨Excited to announce a new post-doctoral position in my lab, focussed on the neural capacity limits of language processing using magnetoencephalography! Apply here: wp.nyu.edu/neurolinglab...
wp.nyu.edu
Postdoctoral position – NYU Neurolinguistics Lab
04333
Iban Dlank (b↔︎d) @ibandlank.bsky.social · 27/08/2026
One solution I'm using that meets several of MIT's recommendations for AI-aware education is described in this 10min video: www.youtube.com/watch?v=ylcF.... It works VERY well (but still imperfect), and can be used for joint work on assigned readings, colab notebooks, discussions, and more!
youtube.com
Dr. Idan Blank: ACCOUNTABLE & AUTHENTIC LEARNING IN THE AGE OF UNCHECKED & ARTIFICIAL INTELLIGENCE
YouTube video by CEILS Media
070
Iban Dlank (b↔︎d) @ibandlank.bsky.social · 18/08/2026
Cue the “viewpoint diversity” bros to explain why we should all have a race scientist in our departments (while no one challenges them because they are “senior”)
020
Iban Dlank (b↔︎d) @ibandlank.bsky.social · 13/08/2026
There will also be a symposium in the upcoming Academy of Aphasia about whether the language network is a real thing. Not sure when they'll post the abstracts though. I'll be there to explain what we mean - and, importantly, don't mean - by the term "language network".
031
Iban Dlank (b↔︎d) @ibandlank.bsky.social · 13/08/2026
@evfedorenko.bsky.social, @neuranna.bsky.social, and Tamar Regev's review (www.nature.com/articles/s41...) has 3 commentaries, with replies by the original authors.
nature.com
The language network as a natural kind within the broader landscape of the human brain - Nature Reviews Neuroscience
Many brain areas support complex language processing behaviours. In this Review, Fedorenko et al. disentangle the ‘core’ language system as functionally distinct from the perceptual and motor brain ar...
150
Reposted by Iban Dlank (b↔︎d)
Colin Phillips @colinphillips.bsky.social · 13/08/2026
We are recruiting a lab manager for our new Psycholinguistics Lab in Oxford! Good opportunity for young researchers, e.g., post-bac/masters. Open to applicants from the UK or abroad. Full details here: my.corehr.com/pls/uoxrecru...
Photo of a statue outside Oxford's Sheldonian Theatre
088
Iban Dlank (b↔︎d) @ibandlank.bsky.social · 13/08/2026
p.s.2 If you are unfamiliar with the architecture of Transformer Language Models, I recommend watching this video before you start the notebook
youtube.com
How Transformers Work: A Detailed, Conceptual Explanation (No Coding / Math)
YouTube video by IbanDlank
020
Iban Dlank (b↔︎d) @ibandlank.bsky.social · 13/08/2026
p.s. needless to say, it all relies on minicons by the amazing @kanishka.bsky.social!
110
Iban Dlank (b↔︎d) @ibandlank.bsky.social · 13/08/2026
5/5 No install or local setup—save a copy, open it in Colab and run cells in order. Good for a course, a lab meeting, or self-study. Hopefully it all runs, but apologies for any errors in the code that I didn't catch!
110
Iban Dlank (b↔︎d) @ibandlank.bsky.social · 13/08/2026
4/5 Finally, switch to GPT2, and learn about probing: train a logistic regression model on GPT2's embeddings to predict whether a sentence is grammatical, then test that model on minimal pairs it's never seen. Simplified version of this paper: aclanthology.org/2026.acl-lon...
aclanthology.org
Implicit Representations of Grammaticality in Language Models
Yingshan Susan Wang, Linlu Qiu, Zhaofeng Wu, Roger P. Levy, Yoon Kim. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.
120
Iban Dlank (b↔︎d) @ibandlank.bsky.social · 13/08/2026
3/5 Next: does the geometry of BERT's embeddings track the gradient acceptability of dative sentences ("gave the book to her" vs. "gave her the book")? Compare LM geometry to human ratings (from the DAIS dataset). Simplified version of this paper: aclanthology.org/2025.cxgsnlp...
aclanthology.org
Meaning-infused grammar: Gradient Acceptability Shapes the Geometric Representations of Constructions in LLMs
Supantho Rakshit, Adele E. Goldberg. Proceedings of the Second International Workshop on Construction Grammars and NLP. 2025.
110
Iban Dlank (b↔︎d) @ibandlank.bsky.social · 13/08/2026
2/5 Start with BERT: extract the contextual embedding of the word "bat" in 3 different sentences. Do the embeddings reflect the difference between "bat" the tool and "bat" the animal? Along the way: how tokenization works, how each LM layer behaves, and why you have to correct for anisotropy.
120
Iban Dlank (b↔︎d) @ibandlank.bsky.social · 13/08/2026
New resource: intro workshop on contextual embeddings in Language Models! Free, standalone Google Colab notebook with lots of explanations. Built for psycholing/cogsci folks (and anyone curious about LM internals). No coding required! Just edit-and-run. 🧵 github.com/idanblank/LM...
github.com
LM-embedding-workshop-data/LM_embedding_workshop_notebook_public.ipynb at main · idanblank/LM-embedding-workshop-data
Data for a workshop on using LM contextual embeddings - idanblank/LM-embedding-workshop-data
1325
Reposted by Iban Dlank (b↔︎d)
Matt Goldrick @mattgoldrick.bsky.social · 12/08/2026
Open until filled (🚨🚨apply now!🚨🚨): Postdoc, speech acoustics and major depressive disorder, with me!, Northwestern Linguistics (working with folks in Psychology and Psychiatry). faculty.wcas.northwestern.edu/matt-goldric...
11116
Iban Dlank (b↔︎d) @ibandlank.bsky.social · 10/08/2026
Made my day - my course "Minds in Brains and Machines" (and "Language as a Cognitive Science") had some unexpected real-world value! Here's a text I got this morning from my past student and RA, Ashley Kim.
A text message reading: Idan! How are you doing? I wanted to give a huge thank you because I got hired at my dream company for knowing Marr’s levels of analysis! So much gratitude for how you taught your classes so thoroughly with knowledge I’ll remember forever!
1150
Iban Dlank (b↔︎d) @ibandlank.bsky.social · 30/07/2026
I doubt it, but it’s an empirical question! If it turns out to be true then we’ll learn something interesting imo. But it wasn’t our research question, so I’ll wait for someone else to try it (on models of the same size…).
010
Iban Dlank (b↔︎d) @ibandlank.bsky.social · 29/07/2026
Thank you for the pointer! Will take a look—it may be relevant to other projects as well!
010
Iban Dlank (b↔︎d) @ibandlank.bsky.social · 22/07/2026
Open Mind
0204
Iban Dlank (b↔︎d) @ibandlank.bsky.social · 16/07/2026
Do language models know who did what to whom? Led by grad student @josephdenning.bsky.social, and Hannah Guo (when she was the first undergrad RA in the lab!)
direct.mit.edu
Do Language Models Know Who Did What to Whom?
Abstract. Language models (LMs) are commonly criticized for not “understanding” language. However, many critiques focus on cognitive abilities that, in humans, are distinct from language processing. H...
2173
Iban Dlank (b↔︎d) @ibandlank.bsky.social · 29/06/2026
I recently submitted a manuscript to OpenMind, and the review process was excellent—much better than many other journals I had submitted to! Highly recommended :)
0121
Iban Dlank (b↔︎d) @ibandlank.bsky.social · 26/06/2026
For years I’ve always started with the same word… I was not prepared for this.
A Wordle badge earned by solving the puzzle in one guess
070
Iban Dlank (b↔︎d) @ibandlank.bsky.social · 19/06/2026
Every time a word violates a textbook grammatical rule, a corpus linguist gets their wings. (Lucas Zelnick using a predicative adjective before the noun)
Lucas Zelnick in a stand up show, telling an audience member: “you have an alive ex-wife”. Screenshot of an Instagram reel.
040
Iban Dlank (b↔︎d) @ibandlank.bsky.social · 10/05/2026
And the next step: when someone claims the neuroimaging study has problems / doesn’t replicate, so some scientists start doubting the psychological phenomenon itself despite decades of behavioral work…
051
Iban Dlank (b↔︎d) @ibandlank.bsky.social · 10/05/2026
Thank you so much! I’m honored you used it in your class, and I’m so happy your students found it helpful!
140
Iban Dlank (b↔︎d) @ibandlank.bsky.social · 08/05/2026
Evidence against syntactic encapsulation in “syntax-specialized” attention heads - now with the thread from Thomas!
090
Iban Dlank (b↔︎d) @ibandlank.bsky.social · 24/04/2026
Congratulations!!! 🎉🎉
010
Iban Dlank (b↔︎d) @ibandlank.bsky.social · 10/03/2026
Evidence against syntactic encapsulation in large language models onlinelibrary.wiley.com/doi/10.1111/...
onlinelibrary.wiley.com
Evidence Against Syntactic Encapsulation in Large Language Models
Transformer-based large language models (LLMs) have recently demonstrated exceptional performance in a variety of linguistic tasks. LLMs primarily combine information across words in a sentence using...
0102
Reposted by Iban Dlank (b↔︎d)
Adele Goldberg @adelegoldberg.bsky.social · 27/02/2026
Idan Blank (UCLA, psych) makes the complex intuitive if you want to learn how LLMs work, watch👇 newly posted to YouTube (no ads) www.youtube.com/watch?v=cGMn...
youtu.be
How Transformers Work: A Detailed, Conceptual Explanation (No Coding / Math)
YouTube video by IbanDlank
06815
Iban Dlank (b↔︎d) @ibandlank.bsky.social · 03/08/2025
Day before month before year??? Blasphemy! I rename my students’ files to end with yyyymmdd so that the sorting makes sense.
020
Reposted by Iban Dlank (b↔︎d)
Anna (Anya) Ivanova @neuranna.bsky.social · 28/07/2025
Looking forward to #CogSci2025 ! Find us throughout the conference
0296
Reposted by Iban Dlank (b↔︎d)
Noga Zaslavsky @nogazs.bsky.social · 21/07/2025
📣 I'm looking for a postdoc to join my lab at NYU! Come work with me on a principled, theory-driven approach to studying language, learning, and reasoning, in humans and AI agents. Apply here: apply.interfolio.com/170656 And come chat with me at #CogSci2025 if interested!
14519
Iban Dlank (b↔︎d) @ibandlank.bsky.social · 22/07/2025
Yes, and I replied to the specific part that doesn’t resonate with me.
100
Iban Dlank (b↔︎d) @ibandlank.bsky.social · 22/07/2025
I don’t think banning a super helpful technology in classrooms is useful. For me, it’s better to create assignments that disincentivize its use (under critical views on AI, this should be easy) and actively teach students to use it critically (when & why it’s useful vs. bad).
100
Reposted by Iban Dlank (b↔︎d)
Dr Anna Leshinskaya @annaleshinskaya.bsky.social · 17/07/2025
Thrilled to announce our symposium, Cognitively Inspired Interpretability inn Large Neural Networks, at #CogSci2025 featuring @taylorwwebb.bsky.social, Ellie Pavlick, Jiahai Feng, Gustaw Opielka, ‪‪@claires012345.bsky.social‬, and Idan Blank!
1133
Iban Dlank (b↔︎d) @ibandlank.bsky.social · 12/06/2025
I think there is an audience for cranky. So many good stand up comedians are cranky! And maybe having a beer or two at a bar makes people more receptive to that?
000
Iban Dlank (b↔︎d) @ibandlank.bsky.social · 12/06/2025
You have such interesting data drama stories! They’re like detective novels.
100
Iban Dlank (b↔︎d) @ibandlank.bsky.social · 01/05/2025
Yes, in class would probably be even better! For me that would mean a flipped classroom and I don’t trust my students to watch lectures at home 😳 but if you can make that happen, that would be awesome! Happy to share my materials for setting up pods, etc.
110