Towards Theoretical Understanding of Transformer Test-Time Computing: Investigation on In-Context Linear Regression

• URL: http://arxiv.org/abs/2508.07571v2
• Date: Tue, 19 Aug 2025 11:54:07 GMT
• Title: Towards Theoretical Understanding of Transformer Test-Time Computing: Investigation on In-Context Linear Regression
• Authors: Xingwu Chen, Miao Lu, Beining Wu, Difan Zou,
• Abstract summary: Using more test-time computation during language model inference, such as generating more intermediate thoughts or sampling multiple candidate answers, has proven effective.<n>This paper takes an initial step toward bridging the gap between practical language model inference and theoretical transformer analysis.
• Score: 16.51420987738846
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Using more test-time computation during language model inference, such as generating more intermediate thoughts or sampling multiple candidate answers, has proven effective in significantly improving model performance. This paper takes an initial step toward bridging the gap between practical language model inference and theoretical transformer analysis by incorporating randomness and sampling. We focus on in-context linear regression with continuous/binary coefficients, where our framework simulates language model decoding through noise injection and binary coefficient sampling. Through this framework, we provide detailed analyses of widely adopted inference techniques. Supported by empirical results, our theoretical framework and analysis demonstrate the potential for offering new insights into understanding inference behaviors in real-world language models.

Related papers

• Reasoning as State Transition: A Representational Analysis of Reasoning Evolution in Large Language Models [50.39102836928242]
  We introduce a representational perspective to investigate the dynamics of the model's internal states.<n>We discover that post-training yields only limited improvement in static initial representation quality.
  arXiv  Detail & Related papers  (2026-01-31T15:23:33Z)
• InverseScope: Scalable Activation Inversion for Interpreting Large Language Models [6.841889611296894]
  InverseScope is an assumption-light and scalable framework for interpreting neural activations via input inversion.<n>To address the inefficiency of sampling in high-dimensional spaces, we propose a novel conditional generation architecture.<n>We also introduce a quantitative evaluation protocol that tests interpretability hypotheses using feature consistency rate computed over the sampled inputs.
  arXiv  Detail & Related papers  (2025-06-09T03:59:28Z)
• Studying the Role of Input-Neighbor Overlap in Retrieval-Augmented Language Models Training Efficiency [3.5634988336513587]
  We investigate how varying levels of query-context overlap affect model performance during both training and inference.<n>Our experiments reveal that increased overlap initially has minimal effect, but substantially improves test-time perplexity and model accelerates learning above a critical threshold.
  arXiv  Detail & Related papers  (2025-05-20T12:58:07Z)
• Toward Understanding In-context vs. In-weight Learning [50.24035812301655]
  We identify simplified distributional properties that give rise to the emergence and disappearance of in-context learning.<n>We then extend the study to a full large language model, showing how fine-tuning on various collections of natural language prompts can elicit similar in-context and in-weight learning behaviour.
  arXiv  Detail & Related papers  (2024-10-30T14:09:00Z)
• PRefLexOR: Preference-based Recursive Language Modeling for Exploratory Optimization of Reasoning and Agentic Thinking [0.0]
  PRefLexOR combines preference optimization with concepts from Reinforcement Learning to enable models to self-teach. We focus on applications in biological materials science and demonstrate the method in a variety of case studies.
  arXiv  Detail & Related papers  (2024-10-16T08:46:26Z)
• RAVEN: In-Context Learning with Retrieval-Augmented Encoder-Decoder Language Models [57.12888828853409]
  RAVEN is a model that combines retrieval-augmented masked language modeling and prefix language modeling. Fusion-in-Context Learning enables the model to leverage more in-context examples without requiring additional training. Our work underscores the potential of retrieval-augmented encoder-decoder language models for in-context learning.
  arXiv  Detail & Related papers  (2023-08-15T17:59:18Z)
• Advancing Counterfactual Inference through Nonlinear Quantile Regression [77.28323341329461]
  We propose a framework for efficient and effective counterfactual inference implemented with neural networks. The proposed approach enhances the capacity to generalize estimated counterfactual outcomes to unseen data. Empirical results conducted on multiple datasets offer compelling support for our theoretical assertions.
  arXiv  Detail & Related papers  (2023-06-09T08:30:51Z)
• Constructing Word-Context-Coupled Space Aligned with Associative Knowledge Relations for Interpretable Language Modeling [0.0]
  The black-box structure of the deep neural network in pre-trained language models seriously limits the interpretability of the language modeling process. A Word-Context-Coupled Space (W2CSpace) is proposed by introducing the alignment processing between uninterpretable neural representation and interpretable statistical logic. Our language model can achieve better performance and highly credible interpretable ability compared to related state-of-the-art methods.
  arXiv  Detail & Related papers  (2023-05-19T09:26:02Z)
• Instance-Based Neural Dependency Parsing [56.63500180843504]
  We develop neural models that possess an interpretable inference process for dependency parsing. Our models adopt instance-based inference, where dependency edges are extracted and labeled by comparing them to edges in a training set.
  arXiv  Detail & Related papers  (2021-09-28T05:30:52Z)
• MINIMALIST: Mutual INformatIon Maximization for Amortized Likelihood Inference from Sampled Trajectories [61.3299263929289]
  Simulation-based inference enables learning the parameters of a model even when its likelihood cannot be computed in practice. One class of methods uses data simulated with different parameters to infer an amortized estimator for the likelihood-to-evidence ratio. We show that this approach can be formulated in terms of mutual information between model parameters and simulated data.
  arXiv  Detail & Related papers  (2021-06-03T12:59:16Z)
• An Empirical Investigation of Pre-Trained Transformer Language Models for Open-Domain Dialogue Generation [23.343006562849126]
  We present an empirical investigation of pre-trained Transformer-based auto-regressive language models for the task of open-domain dialogue generation. Training paradigm of pre-training and fine-tuning is employed to conduct learning. Experiments are conducted on the typical single-turn and multi-turn dialogue corpora such as Weibo, Douban, Reddit, DailyDialog, and Persona-Chat.
  arXiv  Detail & Related papers  (2020-03-09T15:20:21Z)

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.