Ben Prystawski

Paper title and authors

Now up as a reviewed @elife.bsky.social preprint: "Continuous developmental changes in word recognition support language learning across early childhood" elifesciences.org/reviewed-pre...

Using data from ~2000 kids ages 1-6, we quantify links between word recognition and early vocabulary growth!

Now out in Cognition, work with the great @gershbrain.bsky.social @tobigerstenberg.bsky.social on formalizing self-handicapping as rational signaling!
📃 authors.elsevier.com/a/1lo8f2Hx2-...

How do we predict what others will do next? 🤔
We look for patterns. But what are the limits of this ability?
In our new paper at CCN 2025 (@cogcompneuro.bsky.social), we explore the computational constraints of human pattern recognition using the classic game of Rock, Paper, Scissors 🗿📄✂️

My final project from grad school is out now in Dev Psych! Mombasa County preschoolers were more accurate on object-based than picture-based vocabulary assessments, whereas Bay Area preschoolers were equally accurate on object-based and picture-based assessments.

psycnet.apa.org/doiLanding?d...

What do representations tell us about a system? Image of a mouse with a scope showing a vector of activity patterns, and a neural network with a vector of unit activity patterns Common analyses of neural representations: Encoding models (relating activity to task features) drawing of an arrow from a trace saying [on_____on____] to a neuron and spike train. Comparing models via neural predictivity: comparing two neural networks by their R^2 to mouse brain activity. RSA: assessing brain-brain or model-brain correspondence using representational dissimilarity matrices

In neuroscience, we often try to understand systems by analyzing their representations — using tools like regression or RSA. But are these analyses biased towards discovering a subset of what a system represents? If you're interested in this question, check out our new commentary! Thread:

Idiosyncratic but not opaque: Linguistic conventions formed in reference games are interpretable by naïve humans and vision–language models Author(s): Boyce, Veronica; Prystawski, Ben; Tan, Alvin Wei Ming; Frank, Michael C. | Abstract: When are in-group linguistic conventions opaque to non-group members (teen slang like "rizz") or general...

When people form conventions in reference games, how easy are they for outsiders to interpret? (for values of "outsider" that include naïve humans and vision-language models) Check out @vboyce.bsky.social's poster today at #CogSci2025 to find out.
paper: escholarship.org/uc/item/16c4...

Scaling up the think-aloud method The think-aloud method, where participants voice their thoughts as they solve a task, is a valuable source of rich data about human reasoning processes. Yet, it has declined in popularity in contempor...

How can we use modern NLP methods to get lots of granular data from think-aloud experiments? Watch @danielwurgaft.bsky.social explain how in the Reasoning session at 4pm this afternoon at #CogSci2025
paper: arxiv.org/abs/2505.23931

Thinking fast, slow, and everywhere in between in humans and language models Author(s): Prystawski, Ben; Goodman, Noah | Abstract: How do humans adapt how they reason to varying circumstances? Prior research has argued that reasoning comes in two types: a fast, intuitive type ...

How do people trade off between speed and accuracy in reasoning tasks without easy heuristics? Come to my talk, "Thinking fast, slow, and everywhere in between in humans and language models," in the Reasoning session this afternoon #CogSci2025 to find out!
paper: escholarship.org/uc/item/5td9...

🚨New paper! We know models learn distinct in-context learning strategies, but *why*? Why generalize instead of memorize to lower loss? And why is generalization transient?

Our work explains this & *predicts Transformer behavior throughout training* without its weights! 🧵

1/

How can we combine the process-level insight that think-aloud studies give us with the large scale that modern online experiments permit? In our new CogSci paper, we show that speech-to-text models and LLMs enable us to scale up the think-aloud method to large experiments!

Delighted to announce our CogSci '25 workshop at the interface between cognitive science and design 🧠🖌️!

We're calling it: 🏺Minds in the Making🏺
🔗 minds-making.github.io

June – July 2024, free & open to the public
(all career stages, all disciplines)

figure 2 from our preprint, reporting the results from two experiments we measure moral judgments about dividing money between two parties and manipulate the degree of asymmetry in the outside options each party has we find that moral judgments track predictions from rational bargaining models like the nash bargaining solution and the kalai-smorodinsky solution in a negotiation context by contrast, in a donation context, moral intuitions completely reverse, instead tracking redistributive and egalitarian principles preprint link: https://osf.io/preprints/psyarxiv/3uqks_v1

the functional form of moral judgment is (sometimes) the nash bargaining solution

new preprint👇

Despite the world being on fire, I can't help but be thrilled to announce that I'll be starting as an Assistant Professor in the Cognitive Science Program at Dartmouth in Fall '26. I'll be recruiting grad students this upcoming cycle—get in touch if you're interested!

title of paper (in text) plus author list

Time course of word recognition for kids at different ages.

Super excited to submit a big sabbatical project this year: "Continuous developmental changes in word
recognition support language learning across early
childhood": osf.io/preprints/ps...

Hello bluesky world :) excited to share a new paper on data visualization literacy 📈 🧠 w/ @judithfan.bsky.social, @arnavverma.bsky.social, Holly Huey, Hannah Lloyd, @lacepadilla.bsky.social!

📝 preprint: osf.io/preprints/ps...
💻 code: github.com/cogtoolslab/...

Figure 1. A schematic depiction of a model-mechanism mapping between a human learning system (left side) and a cognitive model (right side). Candidate model mechanism mappings are pictured as mapping between representations but also can be in terms of input data, architecture, or learning objective.

Figure 2. Data efficiency in human learning. (left) Order of magnitude of LLM vs. human training data, plotted by human age. Ranges are approximated from Frank (2023a). (right) A schematic depiction of evaluation scaling curves for human learners vs. models plotted by training data quantity.

Paper abstract

AI models are fascinating, impressive, and sometimes problematic. But what can they tell us about the human mind?

In a new review paper, @noahdgoodman.bsky.social and I discuss how modern AI can be used for cognitive modeling: osf.io/preprints/ps...

1/13 New Paper!! We try to understand why some LMs self-improve their reasoning while others hit a wall. The key? Cognitive behaviors! Read our paper on how the right cognitive behaviors can make all the difference in a model's ability to improve with RL! 🧵

New paper in Psychological Review!

In "Causation, Meaning, and Communication" Ari Beller (cicl.stanford.edu/member/ari_b...) develops a computational model of how people use & understand expressions like "caused", "enabled", and "affected".

📃 osf.io/preprints/ps...
📎 github.com/cicl-stanfor...
🧵

The broader spectrum of in-context learning The ability of language models to learn a task from a few examples in context has generated substantial interest. Here, we provide a perspective that situates this type of supervised few-shot learning...

What counts as in-context learning (ICL)? Typically, you might think of it as learning a task from a few examples. However, we’ve just written a perspective (arxiv.org/abs/2412.03782) suggesting interpreting a much broader spectrum of behaviors as ICL! Quick summary thread: 1/7

Hey! Could you add me?

Aerial picture of the UBC campus, with an arrow pointing to a building and text asking "Your PhD lab?"

Do you want to understand how language models work, and how they can change language science? I'm recruiting PhD students at UBC Linguistics! The research will be fun, and Vancouver is lovely. So much cool NLP happening at UBC across both Ling and CS! linguistics.ubc.ca/graduate/adm...

If you try to replicate a finding so you can build on it, but your study fails, what should you do? Should you follow up and try to "rescue" the failed rep, or should you move on? Boyce et al. tried to answer this question; in our sample, 5 of 17 rescue projects succeeded.

osf.io/preprints/ps...

A promotional image of One Hour One Life, showing a character growing up from a baby, to a child, to an adult, to an old woman, to a pile of bones. This work is not affiliated with One Hour One Life; we are grateful to Jason Rohrer, the game's developer, for making the game open data and open source.

Preprint alert! After 4 years, I’m super excited to share work with @thecharleywu.bsky.social @gershbrain.bsky.social and Eric Schulz on the rise and fall of technological development in virtual communities in #OneHourOneLife #ohol
doi.org/10.31234/osf...

Clear clusters in model representations driven by some features (plot colors) but neglecting other more complex ones (plotted as shapes) which are mixed within the color clusters.

How well can we understand an LLM by interpreting its representations? What can we learn by comparing brain and model representations? Our new paper highlights intriguing biases in learned feature representations that make interpreting them more challenging! 1/

When a replication fails, researchers have to decide whether to make another attempt or move on. How should we think about this decision? Here's a new paper trying to answer this question, led by Veronica Boyce and featuring student authors from my class!

osf.io/preprints/ps...