1 citations · 3 across the 11 of their papers we have counts for
13 papers
Polar probe linearly decodes semantic structures from LLMs
Pablo J. Diego-Simón, Pierre Orhan, Emmanuel Chemla +2
How do artificial neural networks bind concepts to form complex semantic structures? Here, we propose a simple neural code, whereby the existence and the type of relations between…
DiscoPhon: Benchmarking the Unsupervised Discovery of Phoneme Inventories With Discrete Speech Units
Maxime Poli, Manel Khentout, Angelo Ortiz Tandazo +3
We introduce DiscoPhon, a multilingual benchmark for evaluating unsupervised phoneme discovery from discrete speech units. DiscoPhon covers 6 dev and 6 test languages, chosen to sp…
Biasless Language Models Learn Unnaturally: How LLMs Fail to Distinguish the Possible from the Impossible
Imry Ziv, Nur Lan, Emmanuel Chemla
Are large language models (LLMs) sensitive to the distinction between humanly possible and impossible languages? This question was recently used in a broader debate on whether LLMs…
Probing Syntax in Large Language Models: Successes and Remaining Challenges
Pablo J. Diego-Simón, Emmanuel Chemla, Jean-Rémi King +1
The syntactic structures of sentences can be readily read-out from the activations of large language models (LLMs). However, the ``structural probes'' that have been developed to r…
A Neural Model for Word Repetition
Daniel Dager, Robin Sobczyk, Emmanuel Chemla +1
It takes several years for the developing brain of a baby to fully master word repetition-the task of hearing a word and repeating it aloud. Repeating a new word, such as from a ne…
A Minimum Description Length Approach to Regularization in Neural Networks
Matan Abudy, Orr Well, Emmanuel Chemla +2
State-of-the-art neural networks can be trained to become remarkable solutions to many problems. But while these architectures can express symbolic, perfect solutions, trained mode…
Large Language Models as Proxies for Theories of Human Linguistic Cognition
Imry Ziv, Nur Lan, Emmanuel Chemla +1
We consider the possible role of current large language models (LLMs) in the study of human linguistic cognition. We focus on the use of such models as proxies for theories of cogn…
Disentanglement and Compositionality of Letter Identity and Letter Position in Variational Auto-Encoder Vision Models
Bruno Bianchi, Aakash Agrawal, Stanislas Dehaene +2
Human readers can accurately count how many letters are in a word (e.g., 7 in ``buffalo''), remove a letter from a given position (e.g., ``bufflo'') or add a new one. The human bra…
A polar coordinate system represents syntax in large language models
Pablo Diego-Simón, Stéphane D'Ascoli, Emmanuel Chemla +2
Originally formalized with symbolic representations, syntactic trees may also be effectively represented in the activations of large language models (LLMs). Indeed, a 'Structural P…
What Makes Two Language Models Think Alike?
Louis Jalouzot, Christophe Pallier, Emmanuel Chemla +1
Do architectural and training differences influence the way models represent and process language? Traditional similarity metrics tell us whether two models share a similar represe…
Metric Learning Encoding Models: A Multivariate Framework for Interpreting Neural Representations
Louis Jalouzot, Christophe Pallier, Emmanuel Chemla +1
Understanding how explicit theoretical features are encoded in opaque neural systems is a central challenge now common to neuroscience and AI. We introduce Metric Learning Encoding…
Bridging the Empirical-Theoretical Gap in Neural Network Formal Language Learning Using Minimum Description Length
Nur Lan, Emmanuel Chemla, Roni Katzir
Neural networks offer good approximation to many tasks but consistently fail to reach perfect generalization, even when theoretical work shows that such perfect solutions can be ex…
Minimum Description Length Hopfield Networks
Matan Abudy, Nur Lan, Emmanuel Chemla +1
Associative memory architectures are designed for memorization but also offer, through their retrieval method, a form of generalization to unseen inputs: stored memories can be see…