Netherlands

Large Language Models have taken the cognitive science world by storm. It is perhaps timely now to take stock of the various research paradigms that have been used to make scientific inferences about &quot;cognition&quot; in these models, human cognition, or learning more generally. We define several non-mutually-exclusive &quot;metaphors&quot; for these research paradigms---GPT-ology, Alien Intelligences, and Existence Proofs---and review recent papers that have studied LLMs under these paradigms. In doing so, we discuss their claims as well as challenges to scientific inference under these various paradigms. We highlight several outstanding issues about LLMs that have to be addressed to push our science forward: closed-source vs open-sourced models; forming conventions on new task &quot;hyperparameters&quot; like instructions and prompts; reproducibility; and (the lack of visibility of) training data.

**Authors:**

Desmond Ong: The University of Texas at Austin

CogSci 2024

GPT-ology, Alien Intelligences, Existence Proofs: How should we think about LLMs in Cognitive Science?

It is my great pleasure and honor to welcome you to the 64th Annual Meeting of the Association for Computational Linguistics (ACL 2026), held in San Diego, California, USA, July 2–7, 2026. ACL 2026 continues our fields tradition of excellence in scholarship, innovation, and inclusivity, and I am deeply grateful to the many volunteers who have worked tirelessly to bring this event to life.

It is exciting to see the many developments in the field reflected at our conference, such as the advancement of large language models (born out of our community) into many application directions. This is not without challenges, partly due to unsolved problems, and partly due to misuse of the technology. All this is reflected well at ACL 2026. The conference also reflects that the field is not only about advancing technology, but also the science of fundamentally understanding language and how it is use by many different speaker communities. <br>
[Full Message from General Chair Philipp Koehn](https://drive.google.com/file/d/1QiJFOmMjvDgsDo3i1XzlQnImxe3RU_Cy/view?usp=sharing) <br><br><br>
<html><button style="display: inline-flex; align-items: center; justify-content: center; white-space: nowrap; border-radius: 9999px; font-weight: bold; background: #7c3aed; color: white; font-family: 'Space Grotesk', sans-serif; height: 40px; font-size: 16px; padding: 0 20px; border: none; cursor: pointer" onclick="window.open('https://drive.google.com/file/d/1QiJFOmMjvDgsDo3i1XzlQnImxe3RU_Cy/view?usp=sharing')">Conference Handbook</button></html>
<br><br><br><br>

To access this event page, you need to log in with the **email address you registered with**. <br>Access credentials will be sent to your email from Underline -  subject line "Welcome to ACL 2026". Please be sure to check your spam email folder if you do not see an email confirmation right away.

To access this event page, you are required to register.
Please complete your registration to continue.

We recommend reading [**the registration information**](https://2026.aclweb.org/registration/) first.

**Online Registration Form**: https://acl.swoogo.com/acl2026

Registration Required

ACL 2026

The 64th Annual Meeting of the Association for Computational Linguistics (ACL 2026) will take place in San Diego, California, from July 2 - 7, 2026.

Large Language Models have taken the cognitive science world by storm. It is perhaps timely now to take stock of the various research paradigms that have been used to make scientific inferences about "cognition" in these models, human cognition, or learning more generally. We define several non-mutually-exclusive "metaphors" for these research paradigms---GPT-ology, Alien Intelligences, and Existence Proofs---and review recent papers that have studied LLMs under these paradigms. In doing so, we discuss their claims as well as challenges to scientific inference under these various paradigms. We highlight several outstanding issues about LLMs that have to be addressed to push our science forward: closed-source vs open-sourced models; forming conventions on new task "hyperparameters" like instructions and prompts; reproducibility; and (the lack of visibility of) training data.

**Authors:**

Desmond Ong: The University of Texas at Austin

poster

The 46th Annual Meeting of the Cognitive Science Society was an in-person meeting held in Rotterdam, The Netherlands at the Postillion Hotel & Conference Centre.

**ON-DEMAND PROGRAM ACCESS AFTER THE CONFERENCE**<br>
Recordings of the invited program are now available. You can access them by clicking on the 'Schedule' icon on the left. Select view by week and navigate to the day and time slot of the recording you wish to view. Click on the time slot to access the recording.

Recordings available (click on each recording to view):

**Thursday, July 25**<br>
0900: [Keynote Speaker Morgan Barense on Enhancing real-world event memory](https://underline.io/events/465/sessions/17997/lecture/99084-dynamics-between-minds-and-the-environment)<br>
1000: [Gleitman Award Winner Isabelle Dautriche on Language Foundations: Insights from acquisition, communication, cognition, and more](https://underline.io/events/465/sessions/18000/lecture/99085-gleitman-talk)<br>
1415: [Invited Symposium: Dynamics between minds and the environment](https://underline.io/events/465/sessions?eventSessionId=18016&searchGroup=lecture)<br>
1700: [Rumelhart Prize Presentation Speaker Alison Gopnik on Exploit, explore, empower: Three ages and three intelligences](https://underline.io/events/465/sessions/18028/lecture/99161-rp1-rumelhart-prize-presentation)<br>

**Friday, July 26**<br>
0900: [C.L. de Carvalho-Heineken Prize Keynote Speaker Kia Nobre on Focusing in memory](https://underline.io/events/465/sessions/18035/lecture/99162-hp1-c-l-de-carvalho-heineken-prize-keynote-address)<br>
1030: [Rumelhart Symposium: Childhood as exploration](https://underline.io/events/465/sessions/18038/lecture/100346-childhood-as-exploration)<br>
1415: [Elman Prize Symposium](https://underline.io/events/465/sessions?eventSessionId=18053&searchGroup=lecture)<br>
1600: [Invited Symposium: Dynamics between minds](https://underline.io/events/465/sessions?eventSessionId=18065&searchGroup=lecture)<br>
1745: [Keynote Speaker Andrea E. Martin on Neural dynamics encode the structure and statistics of language](https://underline.io/events/465/sessions/18077/lecture/99271-neural-dynamics-encode-the-structure-and-statistics-of-language)<br>

**Saturday, July 27**<br>
0900: [Keynote Speaker Gregor Schöner on How higher cognition emerges from the dynamics of strongly interacting neural populations](https://underline.io/events/465/sessions/18083/lecture/99273-dynamics-within-the-mind)<br>
1000: [Glushko Talks](https://underline.io/events/465/sessions?eventSessionId=18086&searchGroup=lecture)<br>
1500: [Invited Symposium: Dynamics within minds](https://underline.io/events/465/sessions?eventSessionId=18100&searchGroup=lecture)<br>


#### **CogSci 2024 Program Book**
<div style="position:relative;padding-top:0;width:900px;height:500px;"><iframe style="position:absolute;border:none;width:100%;height:100%;left:0;top:0;" src="https://online.fliphtml5.com/ebtyf/cble/"  seamless="seamless" scrolling="no" frameborder="0" allowtransparency="true" allowfullscreen="true" ></iframe></div>


**The Program Book can be downloaded [here](https://drive.google.com/file/d/1yRVp1cBqRnAbdYi4GXWV_pZeyWNsRquv/view?usp=sharing).**

#### **Letter from the President**

Welcome to the 2024 Meeting of the Cognitive Science Society in Rotterdam! CogSci 2024 brings together a large community of cognitive scientists who have traveled here from around the world, as well as a smaller group of remote presenters. I would like to extend an especially warm welcome to our first-time CogSci meeting attendees, and hope that 
their experience inspires them to join our future meetings for many years to come.

I want to thank this year’s conference co-chairs, Larissa K Samuelson, Stefan Frank, Mariya Toneva, Allyson Mackey, and Eliot Hazeltine, for putting together an amazing series of invited talks and symposia around the theme of Dynamics of Cognition. The co-chairs also deserve credit for handling a record-breaking number of submissions and producing a truly exciting conference program 

Beyond the invited talks and symposia dedicated to this year’s conference theme, I would like to highlight the presentations and events honoring the 
Rumelhart, Elman, Gleitman and C .L . de Carvalho Heineken prize winners, including the Rumelhart reception on Thursday. As you may know, we are recording all of the invited talks and symposia, and all prize-winners’ talks, and will be making this content available after the conference. 

The Cognitive Science Society relies on the volunteer efforts of our Governing Board members who work on matters related to our membership; conference policies; diversity and inclusion, international, and outreach initiatives; prizes and much more. We welcome interest from Society members who would like to become more involved in what we do. The 
Society is deeply grateful to Rick Dale and Andrea Bender, who serve as editors of the two Society journals, Cognitive Science and Topics in Cognitive Science respectively. We especially acknowledge our editors’ generosity in donating their stipends each year to D&I and outreach efforts. 

This year’s Annual Meeting depends critically on the support of Marischal De Armond and his team at Podium Conference Specialists: Sharon Zwack, Cendrine De Vis, Sarah-Kate Burke, and Rachel Waller worked tirelessly to secure the venue and handled a multitude of logistical and organizational issues, from arranging coffee breaks and receptions to making sure that a huge number of posters fits comfortably into the available space. Finally, and importantly, I would like to acknowledge the Cognitive Science Society Executive Officer, Erica Wojcik, for managing the complex and ever expanding sphere of our Society’s activities.

I hope you enjoy this year’s meeting and the many different opportunities to engage with the vibrant community of cognitive scientists gathered here. I also hope you get a chance to explore the great city of Rotterdam with its many attractions, museums, restaurants and scenic views. If you’d like to talk more about our Society, or simply want to say hello, please stop by!

![](https://assets.underline.io/markdown_image/1/image/3c66d0e21fff30e204df1138ffd15173.jpeg)

**Anna Papafragou**  
President, Cognitive Science 
Society 2023-2024

**CODE OF CONDUCT**<br>
By attending the CogSci 2024 Conference, you are required to adhere to the society’s [Code of Conduct](https://drive.google.com/file/d/16-6KkptF0Gn3ZYGJDlpqpwTdImw45Ng0/view?usp=drive_link).

**ABOUT THE COGNITIVE SCIENCE SOCIETY**<br>
The Cognitive Science Society brings together researchers from around the world who hold a common goal: understanding the nature of the human mind. The mission of the Society is to promote Cognitive Science as a discipline, and to foster scientific interchange among researchers in various areas of study, including Artificial Intelligence, Linguistics, Anthropology, Psychology, Neuroscience, Philosophy, and Education.

The Society is a non-profit professional organization and its activities include sponsoring an annual conference and publishing the journals Cognitive Science and TopiCS.
<br><br>

You need to log in with the email address you registered with. Access credentials have been sent to your email. 

Please be sure to check your spam and other email folders if you do not see an email confirmation right away.

Please log in to explore this event.

It looks like you are not registered for this event. 

To access the site please register [**here**](https://cognitivesciencesociety.org/registration/). 

Please register!

The 46th Annual Meeting of the Cognitive Science Society presents the latest research across cognitive science and highlights the theme of Cognition in Context.

This research explores the capabilities of Large Language Models (LLMs) to engage in abstract and concrete thought processes, challenging the common belief that LLMs are incapable of human-like, abstract thinking. Drawing upon the Construal Level Theory (Trope & Liberman, 2010), we demonstrate how prompts tailored for each construal level (abstract versus concrete) influence LLMs’ performance in tasks requiring different cognitive approaches. Our key findings include: 1) LLMs exhibit a statistically significant difference in construal level depending on the prompt conditions, and 2) LLMs display superior performance in tasks aligned with the prompted construal level; sentiment analysis in concrete conditions and natural language inference in abstract conditions. This research contributes to the scientific understanding of LLMs, offering practical insights for their effective use in tasks necessitating diverse cognitive capabilities.

**Authors:**

Seung Joo Yoo: Seoul National University; Sangah Lee: Seoul National University

Large Language Models Show Human-Like Abstract Thinking Patterns: A Construal-Level Perspective

The human cognitive system exhibits remarkable flexibility
and generalization capabilities, partly due to its ability to form
low-dimensional, compositional representations of the envi-
ronment. In contrast, standard neural network architectures
often struggle with abstract reasoning tasks, overfitting, and
require extensive data for training. This paper investigates the
impact of the relational bottleneck—a mechanism that focuses
processing on relations among inputs—on the learning of fac-
torized representations conducive to compositional coding. We
demonstrate that such a bottleneck not only improves general-
ization and learning efficiency but also aligns network perfor-
mance with human-like behavioral biases. Networks trained
with the relational bottleneck developed orthogonal represen-
tations of the datasets latent features, reflecting the factor-
ized structure thought to underlie human cognitive flexibil-
ity. Moreover, the relational network mimics human biases
towards regularity without pre-specified symbolic primitives,
suggesting that the bottleneck fosters the emergence of abstract
representations akin to symbols.

**Authors:**

Declan I. Campbell: Princeton University; Jonathan Cohen: Princeton University

A Relational Inductive Bias for Dimensional Abstraction in Neural Networks

Previous research about the intensity of emotional facial expressions has relied on stimulus sets of morphed facial expressions that have been generated artificially. Ecologically valid open-access facial stimulus sets with varying intensities of multiple different expressions are rare. However, there is a growing need for a validated facial stimulus set that would include multiple levels of intensities. This study aimed to develop and test the psychometric properties of a stimulus set with real facial expressions (8 men and 8 women) with 11 intensity levels for five facial expression categories: angry, happy, neutral, surprised and sad. 52 individuals rated the valence, arousal and intensity of the 656 stimuli. Descriptive statistics, internal consistency of the rating for each stimulus, emotion category, and intensity level were described. The stimuli and summary data are available upon request (https://osf.io/f8ews/).

**Authors:**

Gerly Tamm: Ghent University

Development and Validation of the Facial Expression Intensity Stimulus Set (FEISS)

We present a novel theory and neural process model of visual selective attention to answer long-standing questions in the field of visual attention. We show that the model with fixed parameter values can explain the unexpected efficiency of triple conjunction search (Nordfang & Wolfe, 2014), the influence of a task-irrelevant size singleton on search (Proulx, 2007), and how a third correlated but task-irrelevant feature improves search efficiency (Found, 1998). It also accounts for critical findings in the attention capture literature without the need to introduce different modes (Bacon & Egeth, 1994), signal-suppression (Gaspelin, Leonard, & Luck, 2015; Gaspelin & Luck, 2018; Lien, Ruthruff, & Hauck, 2021) or an attentional window (Theeuwes, 1992, 2023), shedding new light on recent debates.

**Authors:**

Raul Grieben: Ruhr-Universität Bochum; John Spencer: University of East Anglia; Gregor Schöner: Ruhr-Universität Bochum

Visual selective attention: Priority is all you need

Models of bottom-up visual attention such as the ``saliency map'' predict overt gaze under laboratory conditions while subjects view static images or videos while seated. Here, we show that the saliency map model predicts gaze at similar rates even when applied to video from a head-camera as part of a wearable eye-tracking system (Tobii Pro Glasses 2) while subjects drive an automobile or are passively driven while sitting in the front passenger-side seat. The ability of saliency to predict gaze varies depending on the driving task (saliency better predicts passenger gaze) and external conditions (saliency better predicts gaze at night). We further demonstrate that predictive performance is improved when the head-camera video is transformed to retinal coordinates before feeding it to the saliency model.

**Authors:**

Richard E Veale: Kyoto University; Kenji Murase: Mazda Motor Corporation; Masayuki Watanabe: Mazda Motor Corporation; Tadashi Isa: Kyoto University

Visual saliency predicts gaze during real-world driving task

Various methods of dimensionality reduction have been used to apply a quantitative approach to the study of complex skill acquisition. This work builds upon past approaches, offering a comparative analysis of principal component analysis, logistic regression, and linear discriminant analysis to quantify expertise in the domain of competitive video gaming, or “eSports.” We present a novel, robust dataset of expert and non-expert gameplay data from professional and amateur players of the Super Smash Bros. Melee competitive fighting game. We assess each quantitative model via the metrics of providing accurate expertise classification, predictive utility, and a pragmatic window into the features of complex skill performance that hold the most weight in overall performance outcomes, thereby also providing insights for direction of future training. We posit that linear discriminant analysis provides the best performance for all relevant metrics. The nuances are discussed here, and suggestions for the field are offered for future study of other complex skill domains.

**Authors:**

Noah Phillips: Rensselaer Polytechnic Institute; Chris R. Sims: Rensselaer Polytechnic Institute

Understanding Expertise in Elite Competitive eSports: A Comparison of Approaches to Scalable Dimensionality Reduction

Everyday reasoning about the behavior of those around us involves accounting for why others act the way they do. With multiple available explanations for somebody’s behavior, how do observers choose the best one? A large body of work in social psychology suggests that people rely heavily on traits to explain behavior rather than external factors. Recent results have called this into question, arguing that people balance traits, mental states, and situational constraints to make sense of others’ actions. In the current work, we hypothesize that people rely on counterfactual simulation to weigh different explanations for others’ behavior. We propose a computational model of this process that makes concrete predictions about when people will prefer to explain events based on the actor’s traits or their situation. We test the predictions of this model in an experimental paradigm in which trait and situation each guide behavior to varying degrees. We find that the model predicts people’s trait and situation explanations better than heuristic-based alternatives. These results inform longstanding questions about our intuitive psychology and show how social reasoning can be investigated within the broader context of domain general causal inference.

**Authors:**

Erik Brockbank: Stanford University; Justin Yang: Stanford University; Mishika Govil: Stanford University; Judith E. Fan: Stanford University; Tobias Gerstenberg: Stanford University

Without his cookies, he's just a monster: a counterfactual simulation model of social explanation

The question of how to best and robustly detect human systemic cognitive states like workload, sense of urgency, mind wandering, interference, and others is an open question. The answer depends both on the employed physiological measurements as well as the trained computational classification models. In this paper, we analyze data from a human driving experiment to explore the validity of eye gaze in assessing different systemic human cognitive states. In addition, we explored the association between different pairs of systemic human cognitive states. Our statistical analyses and the classification results indicated that eye gaze, in particular the percentage change in pupil size (PCPS), is a reliable physiological biomarker in assessing multiple systemic human cognitive states including workload, sense of urgency, and mind wandering. Moreover, interference, generated by integrating secondary events which must be completed timely manner into the ongoing task, can be assessed via participant's response times.

**Authors:**

Ayca Aygun: Tufts University; Thuan Nguyen: Worcester Polytechnic Institute; Matthias Scheutz: Tufts University

Assessment of Multiple Systemic Human Cognitive States using Eye Gaze

What structures children’s early language environment? Large corpora of child-centered naturalistic recordings provide an important window into this question, but most available data centers on young children within the home environment or in lab contexts interacting primarily with a single caregiver. Here, we characterize children’s language experience in a very different kind of environment: the preschool classroom. Children ages 3 – 5 years (N = 26) wore a head-mounted camera in their preschool class, yielding a naturalistic, egocentric view of children’s everyday experience across many classroom activity contexts (e.g., block building, snack time), with >30 hours of video data. Using semi-automatic transcriptions (228,000 words), we find that activity contexts in the preschool classroom vary in both the quality and quantity of the language that children both hear and produce. Together, these findings reinforce prior theories emphasizing the contribution of activity contexts in structuring the variability in children’s early learning environments.

**Authors:**

Robert Z. Sparks: Stanford University; Bria Long: Stanford University; Grace E Keene: Stanford University; Malia J. Perez: Stanford University; Alvin Wei Ming Tan: Stanford University; Virginia A Marchman: Stanford University; Michael C. Frank: Stanford University

Characterizing Contextual Variation in Children’s Preschool Language Environment Using Naturalistic Egocentric Videos

People employ a variety of strategies when solving value-based tasks. Here, we examine such differences in the motivational Go/NoGo task, which orthogonalizes action and valence, using open-source data from 817 participants. Using behavioral modelling and analysis, we identified four distinct clusters of people. The first group appeared highly responsive to the task states valence. Two groups exhibited either varying learning rates for cues signaling rewarding and punishing states or differing sensitivities to rewards and punishments. The fourth group (18% of participants) acted like naïve reinforcement learners, with their responses shaped by outcomes in a manner that was independent of the state information provided by the cues. Ignoring state-independence would corrupt inference about the parameters concerning normal state-dependent-learning. Accommodating such population heterogeneity in behavioral-modeling is crucial, especially when working with participants who experience behavioral challenges, and when using models to generate regressors for examining neural data.

**Authors:**

Azadeh Nazemorroaya: Computational Neuroscience, Max Planck Institute for Biological Cybernetics; Dan Bang: Aarhus University; Peter Dayan: Max Planck Institute for Biological Cybernetics

Premium content

Downloads

Next from CogSci 2024

Large Language Models Show Human-Like Abstract Thinking Patterns: A Construal-Level Perspective

Similar lecture

Fundamental Reasoning Paradigms Induce Out-of-Domain Generalization in Language Models

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES