Geometry of Decision Making in Language Models

• URL: http://arxiv.org/abs/2511.20315v1
• Date: Tue, 25 Nov 2025 13:52:46 GMT
• Title: Geometry of Decision Making in Language Models
• Authors: Abhinav Joshi, Divyanshu Bhatt, Ashutosh Modi,
• Abstract summary: Large Language Models (LLMs) show strong generalization across diverse tasks, yet the internal decision-making processes behind their predictions remain opaque.<n>We study the geometry of hidden representations in LLMs through the lens of textitintrinsic dimension (ID)<n>We perform a large-scale study, with 28 open-weight transformer models and estimate ID across layers using multiple estimators.
• Score: 19.74354232642455
• License: http://creativecommons.org/licenses/by-sa/4.0/
• Abstract: Large Language Models (LLMs) show strong generalization across diverse tasks, yet the internal decision-making processes behind their predictions remain opaque. In this work, we study the geometry of hidden representations in LLMs through the lens of \textit{intrinsic dimension} (ID), focusing specifically on decision-making dynamics in a multiple-choice question answering (MCQA) setting. We perform a large-scale study, with 28 open-weight transformer models and estimate ID across layers using multiple estimators, while also quantifying per-layer performance on MCQA tasks. Our findings reveal a consistent ID pattern across models: early layers operate on low-dimensional manifolds, middle layers expand this space, and later layers compress it again, converging to decision-relevant representations. Together, these results suggest LLMs implicitly learn to project linguistic inputs onto structured, low-dimensional manifolds aligned with task-specific decisions, providing new geometric insights into how generalization and reasoning emerge in language models.

Related papers

• TangramPuzzle: Evaluating Multimodal Large Language Models with Compositional Spatial Reasoning [104.66714520975837]
  We introduce a geometry-grounded benchmark designed to evaluate compositional spatial reasoning through the lens of the classic Tangram game.<n>We propose the Tangram Construction Expression (TCE), a symbolic geometric framework that grounds tangram assemblies in exact, machine-verifiable coordinate specifications.<n>We conduct extensive evaluation experiments on advanced open-source and proprietary models, revealing an interesting insight: MLLMs tend to prioritize matching the target silhouette while neglecting geometric constraints.
  arXiv  Detail & Related papers  (2026-01-23T07:35:05Z)
• PlaM: Training-Free Plateau-Guided Model Merging for Better Visual Grounding in MLLMs [59.78917775399492]
  Multimodal instruction fine-tuning paradoxically degrades this text's reasoning capability.<n>We propose a training-free framework to mitigate this degradation.
  arXiv  Detail & Related papers  (2026-01-12T15:27:51Z)
• Visualizing LLM Latent Space Geometry Through Dimensionality Reduction [0.0]
  We extract, process, and visualize latent state geometries in Transformer-based language models through dimensionality reduction.<n>We demonstrate experiments on GPT-2 and LLaMa models, where we uncover interesting geometric patterns in latent space.
  arXiv  Detail & Related papers  (2025-11-26T17:11:39Z)
• Shape Happens: Automatic Feature Manifold Discovery in LLMs via Supervised Multi-Dimensional Scaling [43.70541744169649]
  We introduce Supervised Multi-Dimensional Scaling (SMDS), a model-agnostic method to automatically discover feature manifold.<n>We find that different features form various geometric structures such as circles, lines, and clusters.<n>Our findings shed light on the functional role of feature manifold, supporting a model of entity-based reasoning in which LMs encode and transform structured representations.
  arXiv  Detail & Related papers  (2025-10-01T15:30:47Z)
• How Multimodal LLMs Solve Image Tasks: A Lens on Visual Grounding, Task Reasoning, and Answer Decoding [39.342366994703376]
  We introduce a probing framework to analyze how MLLMs process visual and textual inputs across layers.<n>We show that while the overall stage-wise structure remains stable across variations in visual tokenization, instruction tuning data, and pretraining corpus, the specific layer allocation to each stage shifts.
  arXiv  Detail & Related papers  (2025-08-27T21:22:01Z)
• Quantization Meets dLLMs: A Systematic Study of Post-training Quantization for Diffusion LLMs [78.09559830840595]
  We present the first systematic study on quantizing diffusion-based language models.<n>We identify the presence of activation outliers, characterized by abnormally large activation values.<n>We implement state-of-the-art PTQ methods and conduct a comprehensive evaluation.
  arXiv  Detail & Related papers  (2025-08-20T17:59:51Z)
• The Unreasonable Effectiveness of Model Merging for Cross-Lingual Transfer in LLMs [45.08958917457921]
  Large language models (LLMs) still struggle across tasks outside of high-resource languages.<n>In this work, we investigate cross-lingual transfer to lower-resource languages where task-specific post-training data is scarce.
  arXiv  Detail & Related papers  (2025-05-23T20:28:31Z)
• Multimodal Language Models See Better When They Look Shallower [54.5303326937134]
  Multimodal large language models (MLLMs) typically extract visual features from the final layers of a pretrained Vision Transformer (ViT)<n>We present the first comprehensive study of visual layer selection for MLLMs, analyzing representation similarity across ViT layers.<n>We find that while deep layers excel in semantic-rich tasks like OCR, shallow and middle layers significantly outperform them on fine-grained visual tasks.
  arXiv  Detail & Related papers  (2025-04-30T09:07:10Z)
• Multidimensional Consistency Improves Reasoning in Language Models [21.989335720239467]
  We introduce a framework for testing models for answer consistency across multiple input variations.<n>We induce variations in (i) order of shots in prompt, (ii) problem phrasing, and (iii) languages used.<n>Our framework consistently enhances mathematical reasoning performance on both monolingual dataset GSM8K and multilingual dataset MGSM, especially for smaller models.
  arXiv  Detail & Related papers  (2025-03-04T14:41:05Z)
• Understanding the role of FFNs in driving multilingual behaviour in LLMs [0.0]
  In this paper, we conduct an in-depth analysis of the multilingual capabilities of a family of Large Language Models. We introduce novel metrics to probe the model's multilingual behaviour at different layers and shed light on the impact of architectural choices on multilingual processing.
  arXiv  Detail & Related papers  (2024-04-22T03:47:00Z)
• Exploring Concept Depth: How Large Language Models Acquire Knowledge and Concept at Different Layers? [57.04803703952721]
  Large language models (LLMs) have shown remarkable performances across a wide range of tasks.<n>However, the mechanisms by which these models encode tasks of varying complexities remain poorly understood.<n>We introduce the idea of "Concept Depth" to suggest that more complex concepts are typically acquired in deeper layers.
  arXiv  Detail & Related papers  (2024-04-10T14:56:40Z)
• Characterizing Truthfulness in Large Language Model Generations with Local Intrinsic Dimension [63.330262740414646]
  We study how to characterize and predict the truthfulness of texts generated from large language models (LLMs) We suggest investigating internal activations and quantifying LLM's truthfulness using the local intrinsic dimension (LID) of model activations.
  arXiv  Detail & Related papers  (2024-02-28T04:56:21Z)
• Why Lift so Heavy? Slimming Large Language Models by Cutting Off the Layers [9.549646359252346]
  Large Language Models (LLMs) possess outstanding capabilities in addressing various natural language processing (NLP) tasks.<n>The sheer size of these models poses challenges in terms of storage, training and inference due to the inclusion of billions of parameters through layer stacking.<n>We show that even with fewer layers, LLMs maintain similar or better performance levels, particularly in prompt-based fine-tuning for text classification tasks.
  arXiv  Detail & Related papers  (2024-02-18T20:47:10Z)
• 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.