Phd student at @MPI_NL@Donders, working on multimodal semantic representations in 🖥️ and 🧠
Joined May 2019
- Tweets45
- Following768
- Followers190
- Likes986
1/5 Over a decade of comparing deep neural networks to the human brain—but what have we actually learned? Our new @TrendsCognSci Feature Review synthesizes a decade of brain–DNN comparisons, asking what they reveal about brain function across vision and language.
Co-lead with Lea-Maria Schmitt, @emincelik , and with Floris P. de Lange, and @mtoneva1 💫. We had a lot of fun discussing and writing this paper together :). We'd love to hear your feedback!
5/5 Looking ahead, we hope this synthesis helps shape the next decade of research through deeper cross-modal insights, brain data-driven improvements to DNNs, and model-driven neuroscience experimentation. Read the full review with Open Access:
cell.com/trends/cognitive-sc…
4/5 As the field enters its second decade, we argue that combining constructive (building models) and deconstructive (interpreting models) approaches can move brain–DNN comparisons from descriptive alignment toward a theory-constraining scientific tool.
3/5 We organize findings along three shared dimensions—representations, architectures, and objectives—highlighting where DNNs have validated long-standing neuroscientific hypotheses, where they've generated new ones, and where findings from vision and language converge or diverge
2/5 Why bring vision and language together? Although both address closely related problems, research has largely evolved in parallel across the two fields. Bringing them together reveals shared computational principles—and where each is shaped by its own unique constraints.
Dota Tianai Dong retweeted
Good morning #ICLR2026 ☀️
The Re-Align workshop kicks off in 30 minutes! A full day on what we can actually do with representational alignment, from brains to language models to agents.
Schedule in the thread 👇
Dota Tianai Dong retweeted
Check out our talk today @ ICLR 2026 ReAlign!
Title: Dogs aren’t cars (The Dangers of Poor Objectives for Representational Alignment)
Won both tracks of the ICLR 2026 Representational Alignment Challenge (#1 out of 81 and 90 submissions) by exploiting a quirk in the Centered Kernel Alignment algorithm.
github.com/representational-…
Dota Tianai Dong retweeted
I'm excited to be giving an invited talk at the #iclr2026 Re-Align workshop, talking about how we can improve precise instruction following generalization:
9:15 AM Rio time, Monday, April 27
representational-alignment.g…
Woohoo! So proud of you 🥰🎤
Very excited to have been awarded a Google PhD fellowship in NLP for my work on mechanistic interpretability! Big thanks to @Googleorg for the support, as well as to my supervisors @sandropezzelle @boknilev and @ELLISforEurope for all the help along the way.
🤩🤩🤩🙌🙌🙌
@mntssys and I are excited to announce circuit-tracer, a library that makes circuit-finding simple!
Just type in a sentence, and get out a circuit showing (some of) the features your model uses to predict the next token. Try it on @neuronpedia: shorturl.at/SUX2A
Dota Tianai Dong retweeted
🔄✨ Come join us at the Second edition of the Re-Align workshop @iclr_conf! 🚀🧠 The workshop explores the fascinating question of how artificial and biological systems align in their representations of the world. #ReAlign #ICLR2025
Dota Tianai Dong retweeted
Our representational alignment workshop returns to #ICLR2025! Please submit your work on how ML / cogsci / neuro systems represent the world & what shapes these representations 💭🧠🤖
w/ @thisismyhat @dotadong @sucholutsky @lukas_mut @siddsuresh97
🚨Call for Papers🚨
The Re-Align Workshop is coming back to #ICLR2025
Our CfP is finally up! Come share your representational alignment work at our interdisciplinary workshop at
@iclr_conf
representational-alignment.g…
📢Our Call for paper is out !📢 Please consider submitting any relevant work to the Re-Align 2025 @iclr_conf 🏝️
🚨Call for Papers🚨
The Re-Align Workshop is coming back to #ICLR2025
Our CfP is finally up! Come share your representational alignment work at our interdisciplinary workshop at
@iclr_conf
representational-alignment.g…
Dota Tianai Dong retweeted
New paper submission track at #CCN2025 !
So so happy to have @__init_self speaking about her cool work in person! We had a blast with all the questions and discussions! Can’t wait to hear more about the awesome CorText 🤖 🔥!
Had a great time today visiting @DondersInst and presenting my work to the Predictive Processing lab’s AI club. 🤖Thank you for inviting me @DotaDong !
Dota Tianai Dong retweeted
1/n🤖🧠 New paper alert!📢 In "Assessing Episodic Memory in LLMs with Sequence Order Recall Tasks" (arxiv.org/abs/2410.08133) we introduce SORT as the first method to evaluate episodic memory in large language models. Read on to find out what we discovered!🧵
A huge thank you to @lukas_mut for being our last guest speaker before the summer break! We're excited to continue this discussion and can't wait to host you again in the future 🔜🚀
I just had the pleasure to give an invited talk at @DondersInst! Thanks so much for hosting me @DotaDong! The AI club you are organizing seems to be a really cool series of events 🧠🤖
Dota Tianai Dong retweeted
1/
Really excited to share:
Language in Brains, Minds, and Machines w @Nancy_Kanwisher @ev_fedorenko
@AnnualReviews
We survey the insights that language models (LMs) provide on the question of how language is represented and processed in the human brain.
rb.gy/8afztv
