Neurobiology of Language and Bilingualism

Latest papers

• Every Ablation Is a Dose: Counterweights and the Semblance of Self-Repair

  Areeb Ahmad, Pratinav Seth, Vinay Kumar Sankarapu · 2 October 2026

  Ablate a component of a language model, and other components often appear to adjust and compensate. This phenomenon, termed self-repair, has been observed repeatedly, but its mechanism remains unclear. The most systematic study to date concluded that self-repair is noisy and unlikely to have a singl…

• Beyond Linear Concepts: Discovering and Aligning Non-Linear Concept Manifolds in Large Language Models

  Tido Specht, Elias Benedict Krey, Nils Neukirch, Nils Strodthoff · 2 October 2026

  Understanding information processing in large language models (LLMs) requires dissecting the geometric organization of their internal token representations. While existing mechanistic interpretability (MI) methods seek to extract concepts, they are constrained by a strong linearity assumption challe…

• Persistent Depth Ordering amid Shifting Block-Bypass Responses in Language Model Pretraining

  Shengye Tao, Yinzhu Cheng, Haihua Xie · 2 October 2026

  Layer interventions are widely used to probe the internal organization of language models, yet most analyses examine a single training checkpoint even though model representations and computations evolve throughout pretraining. This leaves open which depth-dependent intervention responses reflect pe…

• When Do Attention-Head Ablations Support Causal Claims? Projection-Level Confounds, Floor Effects, and Matched Controls

  Juli Huang · 2 October 2026

  Attention-head ablation, zeroing a head and measuring the resulting change in task performance, is a common method for inferring which components of a language model are causally responsible for a behavior. We show using GPT-2 small that this inference can be fragile unless the intervention semantic…

• Dissonant ballerinas and crafty carrots: a comparative multi-modal analysis of Italian brain rot

  Anca Dinu, Andra-Maria Florescu, Marius Micluta-Campeanu, Stefana-Arina Tabusca, Claudiu Creanga, Andreiana Mihail · 2 October 2026

  This paper presents a comparative multi-modal analysis of Italian and Romanian brain rot memes, investigating the factors that contribute to its appeal and the linguistic and cultural distinctions between the two versions. To conduct this analysis, we introduce a multi-modal brain rot dataset named …

• Contextual trajectory and incremental contextual displacement: Towards using LLMs to understand dynamic, utterance-specific meaning construction

  Grayson Wycliffe Storer, Julia Witte Zimmerman · 2 October 2026

  Transformer-based large language models (LLMs) such as RoBERTa represent text using contextual word embeddings (CWEs), which alter the embeddings associated with each token based on surrounding context. We construct token-wise incremental trajectories by repeatedly recomputing a token's CWE as succe…

• Better Behavioral Prediction, More Faithful Model Ablations? Evidence from Sequential Choice

  Hanbo Xie · 1 October 2026

  Using predictive models to explain cognition requires more than accurate behavioral predictions. Input ablations offer an appealing route: remove information from a model and interpret the resulting performance change as evidence of its importance for behavior. Yet this inference assumes that the mo…

• Reader Proficiency Shapes Layer-wise Surprisal Profiles

  Akio Hayakawa, Horacio Saggion · 1 October 2026

  Reading behaviour varies not only with linguistic input, but also with reader proficiency. In this study, we investigate whether the layer-wise relationship between surprisal from large language models (LLMs) and human gaze behaviour differs across readers with different levels of proficiency and ac…

• Larry Caused the Car to Stop, But the Model Didn't Notice: Transformer Blindness to the M-Heuristic

  Stefania Butnaru, Claudiu Creanga · 1 October 2026

  Modern transformer models excel at capturing semantic relationships through sentence embeddings, yet their ability to perform pragmatic reasoning remains understudied. This paper investigates whether encoder-based transformers such as DeBERTa employ the M-Heuristic (the neo-Gricean principle that ma…

• Cognitive Expert Language Models Better Align with the Corresponding Brain Systems

  Zhivar Sourati, Mengxuan Helen Wu, Nona Ghazizadeh, Jonas Kaplan, Morteza Dehghani, Samuel A. Nastase · 1 October 2026

  Large language models (LLMs) can predict human brain activity across a variety of brain regions during natural language comprehension. Typically, however, LLM-brain alignment is measured using one model for different regions of the brain, and then model performance is summarized across regions. This…

• Language Models Act on Hidden Valence

  Cameron Berg, Caspar Kaiser · 30 September 2026

  Language models describe some internal states as good and others as bad. But whether models have a stake in them is an open question. Simply asking the model is unlikely to be informative. Any answer may be consistent with genuine introspection, superficial pattern-matching, or with fixed scripts le…

• A mechanistic study of language model introspection

  Jiahong Zou, Xiangkun Sun, Lingkai Kong, Tonghan Wang · 30 September 2026

  Large language models (LLMs) can sometimes report perturbations to their internal activations---even when the input provides no evidence that an intervention occurred. How do models detect and localize such internal changes? We study this question using a controlled task that keeps the input text fi…

• Using LMs to Model the Effects of Context and Coreference during Sentence Comprehension

  Kohei Kajikawa, Lin Ai, Tatsuki Kuribayashi, Ethan Gotlieb Wilcox · 30 September 2026

  Language models (LMs) are often used as a tool to model human language processing. Recent studies suggest that severely restricting LMs' context window improves their fit to human psycholinguistic data by simulating human working memory constraints. However, it is possible that this strict memory-de…

• A Polyphonic Conception of AI Understanding

  Matthieu Queloz, Pierre Beckmann · 30 September 2026

  When a doctor, a judge, or an engineer must decide whether to trust an AI model's output, they cannot avoid asking what the model understands. Purely mathematical or statistical descriptions struggle to distinguish trustworthy from untrustworthy outputs without reintroducing the question of AI under…

• Signatures of semantic search in the activations of large language models

  Luke Leckie, Peter M. Todd, Jacob G. Foster · 29 September 2026

  When recalling lists of concepts (e.g., animals) during the semantic fluency task (SFT), both humans and large language models (LLMs) organise their output into clusters of related items (e.g., sea animals) that are punctuated by strategic switches between clusters. In humans, this pattern can be ex…

• Residual Streams Read, Recurrent States Remember: The Global Workspace in Mamba Models

  Wenlong Wang, Fergal Reid · 29 September 2026

  Can the global-workspace account of transformer representations extend to state-space language models? We fit Jacobian lenses to the residual streams and recurrent states of Mamba-1, Mamba-2 and Mamba-3, using the original 1000-prompt recipe. Joint residual--state readouts improve recovery of known …

• Encoded but Not Decoded: Layer-Localized Evidence for a Three-Level Gap in LLM Syntax

  Zhenyan Lu, He Wang, Xiaohui Huang · 25 September 2026

  A language model can fail a syntactic test in two distinct ways: by not encoding the relevant structure, or by encoding it but failing to use it at the output. Behavioral evaluation alone cannot tell these apart. We propose a three-level evaluation framework (behavioral deployment, LM-head readout, …

• Parts-of-Speech as Emergent Categories in SAE Latent Space

  Alessandro Bondielli, Lucia Passaro, Serena Auriemma, Alessandro Lenci · 25 September 2026

  Sparse AutoEncoders (SAEs) offer a promising way to inspect language model representations, but it is still unclear what kind of linguistic structure their latents expose. We use part-of-speech (PoS) categories as a controlled test case to study whether morpho-syntactic information is encoded by ind…

• Grammatical "grandmother neurons" are rare in LLMs

  Linyang He, Nima Mesgarani · 25 September 2026

  Understanding how Large Language Models (LLMs) encode linguistic structures remains a fundamental challenge in interpretability research. While diagnostic classifiers (or "probes") are widely used for this task, they face significant methodological criticism: training auxiliary classifiers introduce…

• Brain-to-Language Decoding: Tasks, Signals, Methods, Evaluation, Practical Use and Beyond

  Yiqian Yang, Yiqun Duan, Chenyu Liu, Yiqi Wang, Xinliang Zhou, Chin-Teng Lin, Yu Zhang · 24 September 2026

  Brain-to-language decoding translates neural activity associated with language production, internal speech and perception into linguistic or expressive outputs. It offers a route to restoring communication after speech loss and a means of studying how the brain represents language. Advances in neura…

• Comparing Latent Concept Formation in State Space Models and Transformers via Sparse Autoencoders

  Rithin Nagaraj, Rupa Laalasa Oruganti, Prerna Subhashchandra Kunder, Ashwini M Joshi · 22 September 2026

  The quadratic scaling of Transformer self-attention has driven the adoption of sub-quadratic Selective State Space Models (SSMs) like Mamba, which compress past context into a fixed-size recurrent hidden state. This strict informational bottleneck raises a foundational question for mechanistic inter…

• The Answer-Basin Representation Hypothesis: We Are Not Probing or Steering Concepts

  Manjiang Yu, Hongji Li, Zihan Wang, Junwei Chen, Xue Li, Priyanka Singh, Yang Cao, Lijie Hu · 22 September 2026

  The Linear Representation Hypothesis associates high-level concepts with directions in language models, but it remains unclear how these concept-related linear structures are organized within the model. We propose the Answer-Basin Representation Hypothesis: the probability measure induced over answe…

• Do LLMs Choose Like Humans? Using Cognitive Theory to Evaluate LLM Decision-Making

  Johnathan Sun, Andrei Shleifer, Yonatan Belinkov · 22 September 2026

  Large language models (LLMs) exhibit a range of human-like decision-making behaviors, but whether these reflect similar underlying mechanisms or surface-level mimicry remains unclear. We evaluate whether LLM context sensitivity aligns with a cognitive economic theory that explains human behavior thr…

• Read-Best Is Not Steer-Best: A Probing--Steering Layer Dissociation in Omni-Modal Large Language Models

  Yibo Wang, Jisheng Dang, Bimei Wang, Yitao Wu, Wencan Zhang, Hong Peng, Jizhao Liu, Bin Hu, Qi Tian, Tat-Seng Chua · 22 September 2026

  Omni-modal large language models integrate text, audio, and image signals into a shared residual stream, where concepts such as emotion can be linearly decoded and causally modified by activation steering. A common but rarely tested assumption is that the layer with the highest probing accuracy is a…

• Analysing the Linearity of Linguistic Relations in Language Model Embedding Spaces

  Vasudevan Nedumpozhimana, Fathima Thekkekara, John Kelleher · 21 September 2026

  We propose a framework to analyse how strongly different linguistic relations are linearly encoded in language model embedding spaces. We formalise linear encoding via a constrained linear approximation over related and unrelated word pairs and apply this to an extended BATS dataset covering inflect…