Interpreting Multi-Attribute Confounding through Numerical Attributes in Large Language Models

• URL: http://arxiv.org/abs/2511.04053v2
• Date: Mon, 10 Nov 2025 13:39:09 GMT
• Title: Interpreting Multi-Attribute Confounding through Numerical Attributes in Large Language Models
• Authors: Hirohane Takagi, Gouki Minegishi, Shota Kizawa, Issey Sukeda, Hitomi Yanaka,
• Abstract summary: We show that large language models (LLMs) encode real-world numerical correlations but tend to systematically amplify them.<n> irrelevant context induces consistent shifts in magnitude representations, with downstream effects that vary by model size.<n>These findings lay the groundwork for fairer, representation-aware control under multi-attribute entanglement.
• Score: 13.5805504750573
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Although behavioral studies have documented numerical reasoning errors in large language models (LLMs), the underlying representational mechanisms remain unclear. We hypothesize that numerical attributes occupy shared latent subspaces and investigate two questions:(1) How do LLMs internally integrate multiple numerical attributes of a single entity? (2)How does irrelevant numerical context perturb these representations and their downstream outputs? To address these questions, we combine linear probing with partial correlation analysis and prompt-based vulnerability tests across models of varying sizes. Our results show that LLMs encode real-world numerical correlations but tend to systematically amplify them. Moreover, irrelevant context induces consistent shifts in magnitude representations, with downstream effects that vary by model size. These findings reveal a vulnerability in LLM decision-making and lay the groundwork for fairer, representation-aware control under multi-attribute entanglement.

Related papers

• Farther the Shift, Sparser the Representation: Analyzing OOD Mechanisms in LLMs [100.02824137397464]
  We investigate how Large Language Models adapt their internal representations when encountering inputs of increasing difficulty.<n>We reveal a consistent and quantifiable phenomenon: as task difficulty increases, the last hidden states of LLMs become substantially sparser.<n>This sparsity--difficulty relation is observable across diverse models and domains.
  arXiv  Detail & Related papers  (2026-03-03T18:48:15Z)
• Unravelling the Mechanisms of Manipulating Numbers in Language Models [9.583581545538479]
  We explore how language models manipulate numbers and quantify the lower bounds of accuracy of these mechanisms.<n>We find that despite surfacing errors, different language models learn interchangeable representations of numbers that are systematic, highly accurate and universal.<n>Our results lay a fundamental understanding of how pre-trained LLMs manipulate numbers and outline the potential of more accurate probing techniques.
  arXiv  Detail & Related papers  (2025-10-30T09:08:50Z)
• Explaining multimodal LLMs via intra-modal token interactions [55.27436637894534]
  Multimodal Large Language Models (MLLMs) have achieved remarkable success across diverse vision-language tasks, yet their internal decision-making mechanisms remain insufficiently understood.<n>We propose enhancing interpretability by leveraging intra-modal interaction.
  arXiv  Detail & Related papers  (2025-09-26T14:39:13Z)
• Factual Self-Awareness in Language Models: Representation, Robustness, and Scaling [56.26834106704781]
  Factual incorrectness in generated content is one of the primary concerns in ubiquitous deployment of large language models (LLMs)<n>We provide evidence supporting the presence of LLMs' internal compass that dictate the correctness of factual recall at the time of generation.<n>Scaling experiments across model sizes and training dynamics highlight that self-awareness emerges rapidly during training and peaks in intermediate layers.
  arXiv  Detail & Related papers  (2025-05-27T16:24:02Z)
• When can isotropy help adapt LLMs' next word prediction to numerical domains? [53.98633183204453]
  It is shown that the isotropic property of LLM embeddings in contextual embedding space preserves the underlying structure of representations.<n> Experiments show that different characteristics of numerical data and model architectures have different impacts on isotropy.
  arXiv  Detail & Related papers  (2025-05-22T05:10:34Z)
• A Comparative Study of Learning Paradigms in Large Language Models via Intrinsic Dimension [16.671316494925346]
  This study investigates the effects of supervised fine-tuning and in-context learning on the hidden representations of Large Language Models (LLMs)<n>We first explore how the ID of LLM representations evolves during SFT and how it varies due to the number of demonstrations in ICL.<n>We then compare the IDs induced by SFT and ICL and find that ICL consistently induces a higher ID compared to SFT.
  arXiv  Detail & Related papers  (2024-12-09T06:37:35Z)
• Can adversarial attacks by large language models be attributed? [1.2289361708127877]
  We show that certain classes of Large Language Models (LLMs) are not identifiable from their outputs alone.<n>We quantify the explosion in the number of plausible model origins for a given output in recent years.
  arXiv  Detail & Related papers  (2024-11-12T18:28:57Z)
• Graph-based Unsupervised Disentangled Representation Learning via Multimodal Large Language Models [42.17166746027585]
  We introduce a bidirectional weighted graph-based framework to learn factorized attributes and their interrelations within complex data. Specifically, we propose a $beta$-VAE based module to extract factors as the initial nodes of the graph. By integrating these complementary modules, our model successfully achieves fine-grained, practical and unsupervised disentanglement.
  arXiv  Detail & Related papers  (2024-07-26T15:32:21Z)
• MM-SpuBench: Towards Better Understanding of Spurious Biases in Multimodal LLMs [38.93090238335506]
  Spurious bias, a tendency to use spurious correlations between non-essential input attributes and target variables for predictions, has revealed a severe pitfall in deep learning models trained on single modality data. We introduce MM-SpuBench, a comprehensive visual question-answering (VQA) benchmark designed to evaluate MLLMs' reliance on nine distinct categories of spurious correlations. Our findings illuminate the persistence of the reliance on spurious correlations from these models and underscore the urge for new methodologies to mitigate spurious biases.
  arXiv  Detail & Related papers  (2024-06-24T20:29:16Z)
• ReEval: Automatic Hallucination Evaluation for Retrieval-Augmented Large Language Models via Transferable Adversarial Attacks [91.55895047448249]
  This paper presents ReEval, an LLM-based framework using prompt chaining to perturb the original evidence for generating new test cases. We implement ReEval using ChatGPT and evaluate the resulting variants of two popular open-domain QA datasets. Our generated data is human-readable and useful to trigger hallucination in large language models.
  arXiv  Detail & Related papers  (2023-10-19T06:37:32Z)
• A Mechanistic Interpretation of Arithmetic Reasoning in Language Models using Causal Mediation Analysis [128.0532113800092]
  We present a mechanistic interpretation of Transformer-based LMs on arithmetic questions. This provides insights into how information related to arithmetic is processed by LMs.
  arXiv  Detail & Related papers  (2023-05-24T11:43:47Z)
• Competence-Based Analysis of Language Models [21.43498764977656]
  CALM (Competence-based Analysis of Language Models) is designed to investigate LLM competence in the context of specific tasks.<n>We develop a new approach for performing causal probing interventions using gradient-based adversarial attacks.<n>We carry out a case study of CALM using these interventions to analyze and compare LLM competence across a variety of lexical inference tasks.
  arXiv  Detail & Related papers  (2023-03-01T08:53:36Z)
• Examining Scaling and Transfer of Language Model Architectures for Machine Translation [51.69212730675345]
  Language models (LMs) process sequences in a single stack of layers, and encoder-decoder models (EncDec) utilize separate layer stacks for input and output processing. In machine translation, EncDec has long been the favoured approach, but with few studies investigating the performance of LMs.
  arXiv  Detail & Related papers  (2022-02-01T16:20:15Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.