Tootfinder

TokenBlowUp: Resolving Representational Singularities in LLM Token Spaces via Monoidal Transformations
Dongfang Zhao
https://arxiv.org/abs/2507.19747 https://

TokenBlowUp: Resolving Representational Singularities in LLM Token Spaces via Monoidal Transformations
Recent work has provided compelling evidence challenging the foundational manifold hypothesis for the token embedding spaces of Large Language Models (LLMs). These findings reveal the presence of geometric singularities around polysemous tokens, which can lead to representational instability. Existing methodologies, which presuppose a smooth data manifold, are ill-equipped to address such intrinsic structural flaws. In this paper, we formalize this problem in the language of scheme theory and p…

Agent-centric learning: from external reward maximization to internal knowledge curation
Hanqi Zhou, Fryderyk Mantiuk, David G. Nagy, Charley M. Wu
https://arxiv.org/abs/2507.22255

Agent-centric learning: from external reward maximization to internal knowledge curation
The pursuit of general intelligence has traditionally centered on external objectives: an agent's control over its environments or mastery of specific tasks. This external focus, however, can produce specialized agents that lack adaptability. We propose representational empowerment, a new perspective towards a truly agent-centric learning paradigm by moving the locus of control inward. This objective measures an agent's ability to controllably maintain and diversify its own knowledge structures…

Turning the Spell Around: Lightweight Alignment Amplification via Rank-One Safety Injection
Harethah Abu Shairah, Hasan Abed Al Kader Hammoud, George Turkiyyah, Bernard Ghanem
https://arxiv.org/abs/2508.20766

Turning the Spell Around: Lightweight Alignment Amplification via Rank-One Safety Injection
Safety alignment in Large Language Models (LLMs) often involves mediating internal representations to refuse harmful requests. Recent research has demonstrated that these safety mechanisms can be bypassed by ablating or removing specific representational directions within the model. In this paper, we propose the opposite approach: Rank-One Safety Injection (ROSI), a white-box method that amplifies a model's safety alignment by permanently steering its activations toward the refusal-mediating su…

Wider than the sky: an alternative to "mapping" the world onto the brain #implantable

Domain Adaptation via Feature Refinement
Savvas Karatsiolis, Andreas Kamilaris
https://arxiv.org/abs/2508.16124 https://arxiv.org/pdf/2508.16124

Domain Adaptation via Feature Refinement
We propose Domain Adaptation via Feature Refinement (DAFR2), a simple yet effective framework for unsupervised domain adaptation under distribution shift. The proposed method synergistically combines three key components: adaptation of Batch Normalization statistics using unlabeled target data, feature distillation from a source-trained model and hypothesis transfer. By aligning feature distributions at the statistical and representational levels, DAFR2 produces robust and domain-invariant feat…

Replaced article(s) found for q-bio.NC. https://arxiv.org/list/q-bio.NC/new
[1/1]:
- In silico discovery of representational relationships across visual cortex
Alessandro T. Gifford, Maya A. Jastrz\k{e}bowska, Johannes J. D. Singer, Radoslaw M. Cichy

Recommender systems, representativeness, and online music: A psychosocial analysis of Italian listeners
Lorenzo Porcaro, Chiara Monaldi
https://arxiv.org/abs/2507.18169 https://…

Recommender systems, representativeness, and online music: A psychosocial analysis of Italian listeners
Recommender systems shape music listening worldwide due to their widespread adoption in online platforms. Growing concerns about representational harms that these systems may cause are nowadays part of the scientific and public debate, wherein music listener perspectives are oftentimes reported and discussed from a cognitive-behaviorism perspective, but rarely contextualised under a psychosocial and cultural lens. We proceed in this direction, by interviewing a group of Italian music listeners …

When Alignment Hurts: Decoupling Representational Spaces in Multilingual Models
Ahmed Elshabrawy, Hour Kaing, Haiyue Song, Alham Fikri Aji, Hideki Tanaka, Masao Utiyama, Raj Dabre
https://arxiv.org/abs/2508.12803

When Alignment Hurts: Decoupling Representational Spaces in Multilingual Models
Alignment with high-resource standard languages is often assumed to aid the modeling of related low-resource varieties. We challenge this assumption by demonstrating that excessive representational entanglement with a dominant variety, such as Modern Standard Arabic (MSA) in relation to Arabic dialects, can actively hinder generative modeling. We present the first comprehensive causal study of this phenomenon by analyzing and directly intervening in the internal representation geometry of large…

Generative Representational Learning of Foundation Models for Recommendation
Zheli Zhou, Chenxu Zhu, Jianghao Lin, Bo Chen, Ruiming Tang, Weinan Zhang, Yong Yu
https://arxiv.org/abs/2506.11999

Generative Representational Learning of Foundation Models for Recommendation
Developing a single foundation model with the capability to excel across diverse tasks has been a long-standing objective in the field of artificial intelligence. As the wave of general-purpose foundation models sweeps across various domains, their influence has significantly extended to the field of recommendation systems. While recent efforts have explored recommendation foundation models for various generative tasks, they often overlook crucial embedding tasks and struggle with the complexit…

Dual-Role Dynamics in Prompting: Elementary Pre-service Teachers' AI Prompting Strategies for Representational Choices
Razan Hamed, Amogh Sirnoorkar, N. Sanjay Rebello
https://arxiv.org/abs/2508.14760

Dual-Role Dynamics in Prompting: Elementary Pre-service Teachers' AI Prompting Strategies for Representational Choices
Pre-service teachers play a unique dual role as they straddle between the roles of students and future teachers. This dual role requires them to adopt both the learner's and the instructor's perspectives while engaging with pedagogical and content knowledge. The current study investigates how pre-service elementary teachers taking a physical science course prompt AI to generate representations that effectively communicate conceptual ideas to two distinct audiences. The context involves particip…

Understanding and Meeting Practitioner Needs When Measuring Representational Harms Caused by LLM-Based Systems
Emma Harvey, Emily Sheng, Su Lin Blodgett, Alexandra Chouldechova, Jean Garcia-Gathright, Alexandra Olteanu, Hanna Wallach
https://arxiv.org/abs/2506.04482

Understanding and Meeting Practitioner Needs When Measuring Representational Harms Caused by LLM-Based Systems
The NLP research community has made publicly available numerous instruments for measuring representational harms caused by large language model (LLM)-based systems. These instruments have taken the form of datasets, metrics, tools, and more. In this paper, we examine the extent to which such instruments meet the needs of practitioners tasked with evaluating LLM-based systems. Via semi-structured interviews with 12 such practitioners, we find that practitioners are often unable to use publicly a…

Generating Human-AI Collaborative Design Sequence for 3D Assets via Differentiable Operation Graph
Xiaoyang Huang, Bingbing Ni, Wenjun Zhang
https://arxiv.org/abs/2508.17645 htt…

Generating Human-AI Collaborative Design Sequence for 3D Assets via Differentiable Operation Graph
The emergence of 3D artificial intelligence-generated content (3D-AIGC) has enabled rapid synthesis of intricate geometries. However, a fundamental disconnect persists between AI-generated content and human-centric design paradigms, rooted in representational incompatibilities: conventional AI frameworks predominantly manipulate meshes or neural representations (\emph{e.g.}, NeRF, Gaussian Splatting), while designers operate within parametric modeling tools. This disconnection diminishes the pr…

Information-computation trade-offs in non-linear transforms
Connor Ding, Abhiram Rao Gorle, Jiwon Jeong, Naomi Sagan, Tsachy Weissman
https://arxiv.org/abs/2506.15948

Information-computation trade-offs in non-linear transforms
In this work, we explore the interplay between information and computation in non-linear transform-based compression for broad classes of modern information-processing tasks. We first investigate two emerging nonlinear data transformation frameworks for image compression: Implicit Neural Representations (INRs) and 2D Gaussian Splatting (GS). We analyze their representational properties, behavior under lossy compression, and convergence dynamics. Our results highlight key trade-offs between INR'…

The harder life gets, the more conventionally melodic, naive and childlike my music gets.
It surprises me, but then I remember when an artist friend became very depressed (not what’s going on with me) her work went from extraordinarily detailed, insectoid abstract ink drawings to much more colourful, simple, “innocent” representational paintings.

Employing Discrete Fourier Transform in Representational Learning
Raoof HojatJalali, Edmondo Trentin
https://arxiv.org/abs/2506.06765 https://

Employing Discrete Fourier Transform in Representational Learning
Image Representation learning via input reconstruction is a common technique in machine learning for generating representations that can be effectively utilized by arbitrary downstream tasks. A well-established approach is using autoencoders to extract latent representations at the network's compression point. These representations are valuable because they retain essential information necessary for reconstructing the original input from the compressed latent space. In this paper, we propose an…

Most #RESTful #APIs aren't really RESTful
https://florian-krae…

How brains build higher order representations of uncertainty
Megan A. K. Peters, Hojjat Azimi Asrari
https://arxiv.org/abs/2506.19057 https://

How brains build higher order representations of uncertainty
Higher-order representations (HORs) are neural or computational states that are "about" first-order representations (FORs), encoding information not about the external world per se but about the agent's own representational processes -- such as the reliability, source, or structure of a FOR. These HORs appear critical to metacognition, learning, and even consciousness by some accounts, yet their dimensionality, construction, and neural substrates remain poorly understood. Here, we propose that …

Foundation Models for Cross-Domain EEG Analysis Application: A Survey
Hongqi Li, Yitong Chen, Yujuan Wang, Weihang Ni, Haodong Zhang
https://arxiv.org/abs/2508.15716 https://

Foundation Models for Cross-Domain EEG Analysis Application: A Survey
Electroencephalography (EEG) analysis stands at the forefront of neuroscience and artificial intelligence research, where foundation models are reshaping the traditional EEG analysis paradigm by leveraging their powerful representational capacity and cross-modal generalization. However, the rapid proliferation of these techniques has led to a fragmented research landscape, characterized by diverse model roles, inconsistent architectures, and a lack of systematic categorization. To bridge this g…

Identifying Algorithmic and Domain-Specific Bias in Parliamentary Debate Summarisation
Eoghan Cunningham, James Cross, Derek Greene
https://arxiv.org/abs/2507.14221

Identifying Algorithmic and Domain-Specific Bias in Parliamentary Debate Summarisation
The automated summarisation of parliamentary debates using large language models (LLMs) offers a promising way to make complex legislative discourse more accessible to the public. However, such summaries must not only be accurate and concise but also equitably represent the views and contributions of all speakers. This paper explores the use of LLMs to summarise plenary debates from the European Parliament and investigates the algorithmic and representational biases that emerge in this context.…

How Causal Abstraction Underpins Computational Explanation
Atticus Geiger, Jacqueline Harding, Thomas Icard
https://arxiv.org/abs/2508.11214 https://arxiv.…

How Causal Abstraction Underpins Computational Explanation
Explanations of cognitive behavior often appeal to computations over representations. What does it take for a system to implement a given computation over suitable representational vehicles within that system? We argue that the language of causality -- and specifically the theory of causal abstraction -- provides a fruitful lens on this topic. Drawing on current discussions in deep learning with artificial neural networks, we illustrate how classical themes in the philosophy of computation and …

Open Questions about Time and Self-reference in Living Systems
Samson Abramsky, Wolfgang Banzhaf, Leo S. D. Caves, Michael Levin, Penousal Machado, Charles Ofria, Susan Stepney, Roger White
https://arxiv.org/abs/2508.11423

Open Questions about Time and Self-reference in Living Systems
Living systems exhibit a range of fundamental characteristics: they are active, self-referential, self-modifying systems. This paper explores how these characteristics create challenges for conventional scientific approaches and why they require new theoretical and formal frameworks. We introduce a distinction between 'natural time', the continuing present of physical processes, and 'representational time', with its framework of past, present and future that emerges with life itself. Representa…

Hecto: Modular Sparse Experts for Adaptive and Interpretable Reasoning
Sanskar Pandey, Ruhaan Chopra, Saad Murtaza Bhat, Ark Abhyudaya
https://arxiv.org/abs/2506.22919

Hecto: Modular Sparse Experts for Adaptive and Interpretable Reasoning
Mixture-of-Experts (MoE) models enable conditional computation by routing inputs to specialized experts, but these experts rely on identical inductive biases, thus limiting representational diversity. This static computation pathway is inefficient for inputs that require different types of reasoning and limits specialization and interpretability. We propose Hecto, a lightweight MoE architecture that leverages architectural heterogeneity by combining a GRU expert for temporal reasoning and an FF…

Value-Set Iteration: Computing Optimal Correlated Equilibria in Infinite-Horizon Multi-Player Stochastic Games
Jiarui Gan, Rupak Majumdar
https://arxiv.org/abs/2506.07186

Value-Set Iteration: Computing Optimal Correlated Equilibria in Infinite-Horizon Multi-Player Stochastic Games
We study the problem of computing optimal correlated equilibria (CEs) in infinite-horizon multi-player stochastic games, where correlation signals are provided over time. In this setting, optimal CEs require history-dependent policies; this poses new representational and algorithmic challenges as the number of possible histories grows exponentially with the number of time steps. We focus on computing $(ε, δ)$-optimal CEs -- solutions that achieve a value within $ε$ of an optimal CE, while al…

Gender-Neutral Machine Translation Strategies in Practice
Hillary Dawkins, Isar Nejadgholi, Chi-kiu Lo
https://arxiv.org/abs/2506.15676 https://

Gender-Neutral Machine Translation Strategies in Practice
Gender-inclusive machine translation (MT) should preserve gender ambiguity in the source to avoid misgendering and representational harms. While gender ambiguity often occurs naturally in notional gender languages such as English, maintaining that gender neutrality in grammatical gender languages is a challenge. Here we assess the sensitivity of 21 MT systems to the need for gender neutrality in response to gender ambiguity in three translation directions of varying difficulty. The specific gen…

Multi-Plasticity Synergy with Adaptive Mechanism Assignment for Training Spiking Neural Networks
Yuzhe Liu, Xin Deng, Qiang Yu
https://arxiv.org/abs/2508.13673 https://

Multi-Plasticity Synergy with Adaptive Mechanism Assignment for Training Spiking Neural Networks
Spiking Neural Networks (SNNs) are promising brain-inspired models known for low power consumption and superior potential for temporal processing, but identifying suitable learning mechanisms remains a challenge. Despite the presence of multiple coexisting learning strategies in the brain, current SNN training methods typically rely on a single form of synaptic plasticity, which limits their adaptability and representational capability. In this paper, we propose a biologically inspired training…

Universal dimensions of visual representation https://www.science.org/doi/10.1126/sciadv.adw7697 "our approach could be extended beyond vision models to examine the representational dimensions that are shared across vision and language";

Emergence of Quantised Representations Isolated to Anisotropic Functions
George Bird
https://arxiv.org/abs/2507.12070 https://arxiv.o…

Emergence of Quantised Representations Isolated to Anisotropic Functions
This paper describes a novel methodology for determining representational alignment, developed upon the existing Spotlight Resonance method. Using this, it is found that algebraic symmetries of network primitives are a strong predictor for task-agnostic structure in representations. Particularly, this new tool is used to gain insight into how discrete representations can form and arrange in autoencoder models, through an ablation study where only the activation function is altered. Representati…

SAM2-UNeXT: An Improved High-Resolution Baseline for Adapting Foundation Models to Downstream Segmentation Tasks
Xinyu Xiong, Zihuang Wu, Lei Zhang, Lei Lu, Ming Li, Guanbin Li
https://arxiv.org/abs/2508.03566

SAM2-UNeXT: An Improved High-Resolution Baseline for Adapting Foundation Models to Downstream Segmentation Tasks
Recent studies have highlighted the potential of adapting the Segment Anything Model (SAM) for various downstream tasks. However, constructing a more powerful and generalizable encoder to further enhance performance remains an open challenge. In this work, we propose SAM2-UNeXT, an advanced framework that builds upon the core principles of SAM2-UNet while extending the representational capacity of SAM2 through the integration of an auxiliary DINOv2 encoder. By incorporating a dual-resolution st…

Convergent transformations of visual representation in brains and models
Pablo Marcos-Manch\'on, Llu\'is Fuentemilla
https://arxiv.org/abs/2507.13941

Convergent transformations of visual representation in brains and models
A fundamental question in cognitive neuroscience is what shapes visual perception: the external world's structure or the brain's internal architecture. Although some perceptual variability can be traced to individual differences, brain responses to naturalistic stimuli evoke similar activity patterns across individuals, suggesting a convergent representational principle. Here, we test if this stimulus-driven convergence follows a common trajectory across people and deep neural networks (DNNs) d…

(YouTube) Dr Zvi Roth on "Representational drift in human visual cortex" https://www.youtube.com/watch?v=5ItxLiJ1M-Y by @…; what does it …

Do We Really Need Specialization? Evaluating Generalist Text Embeddings for Zero-Shot Recommendation and Search
Matteo Attimonelli, Alessandro De Bellis, Claudio Pomo, Dietmar Jannach, Eugenio Di Sciascio, Tommaso Di Noia
https://arxiv.org/abs/2507.05006

Do We Really Need Specialization? Evaluating Generalist Text Embeddings for Zero-Shot Recommendation and Search
Pre-trained language models (PLMs) are widely used to derive semantic representations from item metadata in recommendation and search. In sequential recommendation, PLMs enhance ID-based embeddings through textual metadata, while in product search, they align item characteristics with user intent. Recent studies suggest task and domain-specific fine-tuning are needed to improve representational power. This paper challenges this assumption, showing that Generalist Text Embedding Models (GTEs), p…

Towards Integrated Alignment
Ben Y. Reis, William La Cava
https://arxiv.org/abs/2508.06592 https://arxiv.org/pdf/2508.06592

Towards Integrated Alignment
As AI adoption expands across human society, the problem of aligning AI models to match human preferences remains a grand challenge. Currently, the AI alignment field is deeply divided between behavioral and representational approaches, resulting in narrowly aligned models that are more vulnerable to increasingly deceptive misalignment threats. In the face of this fragmentation, we propose an integrated vision for the future of the field. Drawing on related lessons from immunology and cybersecu…

Kolmogorov-Arnold Wavefunctions
Paulo F. Bedaque, Jacob Cigliano, Hersh Kumar, Srijit Paul, Suryansh Rajawat
https://arxiv.org/abs/2506.02171 https://

Kolmogorov-Arnold Wavefunctions
This work investigates Kolmogorov-Arnold network-based wavefunction ansatz as viable representations for quantum Monte Carlo simulations. Through systematic analysis of one-dimensional model systems, we evaluate their computational efficiency and representational power against established methods. Our numerical experiments suggest some efficient training methods and we explore how the computational cost scales with desired precision, particle number, and system parameters. Roughly speaking, KAN…

Asking #AI models: How will representational drift in human visual cortex affect the need to often recalibrate a Neuralink Blindsight brain implant for restoring vision to the blind?
ChatGPT: https://chatgpt.com/…

ChatGPT - Hochrechnungen vor Wahlschluss
Shared via ChatGPT

Integer Binary-Range Alignment Neuron for Spiking Neural Networks
Binghao Ye, Wenjuan Li, Dong Wang, Man Yao, Bing Li, Weiming Hu, Dong Liang, Kun Shang
https://arxiv.org/abs/2506.05679

Integer Binary-Range Alignment Neuron for Spiking Neural Networks
Spiking Neural Networks (SNNs) are noted for their brain-like computation and energy efficiency, but their performance lags behind Artificial Neural Networks (ANNs) in tasks like image classification and object detection due to the limited representational capacity. To address this, we propose a novel spiking neuron, Integer Binary-Range Alignment Leaky Integrate-and-Fire to exponentially expand the information expression capacity of spiking neurons with only a slight energy increase. This is a…

This https://arxiv.org/abs/2506.00484 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_qbi…

Reconstruction of Partial Dissimilarity Matrices for Cognitive Neuroscience
In cognitive neuroscience research, Representational Dissimilarity Matrices (RDMs) are often incomplete because pairwise similarity judgments cannot always be exhaustively collected as the number of pairs rapidly increases with the number of conditions. Existing methods to fill these missing values, such as deep neural network imputation, are powerful but computationally demanding and relatively opaque. We introduce a simple algorithm based on geometric inference that fills missing dissimilarit…

Reconstruction of Partial Dissimilarity Matrices for Cognitive Neuroscience
Denise Moerel, Tijl Grootswagers
https://arxiv.org/abs/2506.00484 https://

Reconstruction of Partial Dissimilarity Matrices for Cognitive Neuroscience
In cognitive neuroscience research, Representational Dissimilarity Matrices (RDMs) are often incomplete because pairwise similarity judgments cannot always be exhaustively collected as the number of pairs rapidly increases with the number of conditions. Existing methods to fill these missing values, such as deep neural network imputation, are powerful but computationally demanding and relatively opaque. We introduce a simple algorithm based on geometric inference that fills missing dissimilarit…