Understanding the Language Model to Solve the Symbolic Multi-Step Reasoning Problem from the Perspective of Buffer Mechanism

• URL: http://arxiv.org/abs/2405.15302v3
• Date: Tue, 09 Sep 2025 02:51:58 GMT
• Title: Understanding the Language Model to Solve the Symbolic Multi-Step Reasoning Problem from the Perspective of Buffer Mechanism
• Authors: Zhiwei Wang, Yunji Wang, Zhongwang Zhang, Zhangchen Zhou, Hui Jin, Tianyang Hu, Jiacheng Sun, Zhenguo Li, Yaoyu Zhang, Zhi-Qin John Xu,
• Abstract summary: We construct a symbolic multi-step reasoning task to investigate the information propagation mechanisms in Transformer models.<n>We propose a random matrix-based algorithm to enhance the model's reasoning ability.
• Score: 68.05754701230039
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Large language models have consistently struggled with complex reasoning tasks, such as mathematical problem-solving. Investigating the internal reasoning mechanisms of these models can help us design better model architectures and training strategies, ultimately enhancing their reasoning capability. In this study, we constructed a symbolic multi-step reasoning task to investigate the information propagation mechanisms in Transformer models when solving the task through direct answering and Chain-of-Thought (CoT) reasoning. We introduced the concept of buffer mechanism: the model stores various information in distinct buffers and selectively extracts it through the query-key matrix. We proposed a random matrix-based algorithm to enhance the model's reasoning ability. This algorithm introduces only 132 trainable parameters, yet leads to significant performance improvements on 7 multi-step reasoning datasets, including PrOntoQA, LogicAsker, and LogicInference. These findings provide new insights into understanding the large language models.

Related papers

• Fluid Representations in Reasoning Models [91.77876704697779]
  We present a mechanistic analysis of how QwQ-32B processes abstract structural information.<n>We find that QwQ-32B gradually improves its internal representation of actions and concepts during reasoning.
  arXiv  Detail & Related papers  (2026-02-04T18:34:50Z)
• Modeling Hierarchical Thinking in Large Reasoning Models [2.429493364781869]
  Large Language Models (LLMs) have demonstrated remarkable reasoning abilities when they generate step-by-step solutions.<n>When trained to using chain-of-thought reasoning examples, the resulting models appear to learn hierarchical thinking strategies similar to those used by humans.<n>In this paper, we adopt a memoryless Finite State Machine formulation to approximate LRM's emerging hierarchical reasoning dynamics as a structured, interpretable abstraction.
  arXiv  Detail & Related papers  (2025-10-25T21:25:30Z)
• Beyond Memorization: Extending Reasoning Depth with Recurrence, Memory and Test-Time Compute Scaling [60.63703438729223]
  We show how different architectures and training methods affect model multi-step reasoning capabilities.<n>We confirm that increasing model depth plays a crucial role for sequential computations.
  arXiv  Detail & Related papers  (2025-08-22T18:57:08Z)
• AdapThink: Adaptive Thinking Preferences for Reasoning Language Model [32.47427081297578]
  Reinforcement Learning (RL)-based post-training has significantly advanced the complex reasoning capabilities of language models.<n>However, this slow thinking'' paradigm presents a critical challenge to reasoning efficiency.<n>We propose AdapThink, an adaptive post-training framework designed to induce more efficient thinking.
  arXiv  Detail & Related papers  (2025-06-23T02:06:04Z)
• The Reasoning-Memorization Interplay in Language Models Is Mediated by a Single Direction [34.86855316803838]
  We identify a set of linear features in the model's residual stream that govern the balance between genuine reasoning and memory recall.<n>We show that intervening in these reasoning features helps the model more accurately activate the most relevant problem-solving capabilities during answer generation.
  arXiv  Detail & Related papers  (2025-03-29T14:00:44Z)
• A Survey of Model Architectures in Information Retrieval [64.75808744228067]
  We focus on two key aspects: backbone models for feature extraction and end-to-end system architectures for relevance estimation. We trace the development from traditional term-based methods to modern neural approaches, particularly highlighting the impact of transformer-based models and subsequent large language models (LLMs) We conclude by discussing emerging challenges and future directions, including architectural optimizations for performance and scalability, handling of multimodal, multilingual data, and adaptation to novel application domains beyond traditional search paradigms.
  arXiv  Detail & Related papers  (2025-02-20T18:42:58Z)
• Generalized Factor Neural Network Model for High-dimensional Regression [50.554377879576066]
  We tackle the challenges of modeling high-dimensional data sets with latent low-dimensional structures hidden within complex, non-linear, and noisy relationships. Our approach enables a seamless integration of concepts from non-parametric regression, factor models, and neural networks for high-dimensional regression.
  arXiv  Detail & Related papers  (2025-02-16T23:13:55Z)
• Mechanistic Unveiling of Transformer Circuits: Self-Influence as a Key to Model Reasoning [9.795934690403374]
  It is still unclear which multi-step reasoning mechanisms are used by language models to solve such tasks. We employ circuit analysis and self-influence functions to evaluate the changing importance of each token throughout the reasoning process. We demonstrate that the underlying circuits reveal a human-interpretable reasoning process used by the model.
  arXiv  Detail & Related papers  (2025-02-13T07:19:05Z)
• On the Reasoning Capacity of AI Models and How to Quantify It [0.0]
  Large Language Models (LLMs) have intensified the debate surrounding the fundamental nature of their reasoning capabilities. While achieving high performance on benchmarks such as GPQA and MMLU, these models exhibit limitations in more complex reasoning tasks. We propose a novel phenomenological approach that goes beyond traditional accuracy metrics to probe the underlying mechanisms of model behavior.
  arXiv  Detail & Related papers  (2025-01-23T16:58:18Z)
• Cliqueformer: Model-Based Optimization with Structured Transformers [102.55764949282906]
  We develop a model that learns the structure of an MBO task and empirically leads to improved designs. We evaluate Cliqueformer on various tasks, ranging from high-dimensional black-box functions to real-world tasks of chemical and genetic design.
  arXiv  Detail & Related papers  (2024-10-17T00:35:47Z)
• Interpreting token compositionality in LLMs: A robustness analysis [10.777646083061395]
  Constituent-Aware Pooling (CAP) is a methodology designed to analyse how large language models process linguistic structures. CAP intervenes in model activations through constituent-based pooling at various model levels.
  arXiv  Detail & Related papers  (2024-10-16T18:10:50Z)
• Unified Explanations in Machine Learning Models: A Perturbation Approach [0.0]
  Inconsistencies between XAI and modeling techniques can have the undesirable effect of casting doubt upon the efficacy of these explainability approaches. We propose a systematic, perturbation-based analysis against a popular, model-agnostic method in XAI, SHapley Additive exPlanations (Shap) We devise algorithms to generate relative feature importance in settings of dynamic inference amongst a suite of popular machine learning and deep learning methods, and metrics that allow us to quantify how well explanations generated under the static case hold.
  arXiv  Detail & Related papers  (2024-05-30T16:04:35Z)
• Refined Mechanism Design for Approximately Structured Priors via Active Regression [50.71772232237571]
  We consider the problem of a revenue-maximizing seller with a large number of items for sale to $n$ strategic bidders. It is well-known that optimal and even approximately-optimal mechanisms for this setting are notoriously difficult to characterize or compute.
  arXiv  Detail & Related papers  (2023-10-11T20:34:17Z)
• Large Language Models as Analogical Reasoners [155.9617224350088]
  Chain-of-thought (CoT) prompting for language models demonstrates impressive performance across reasoning tasks. We introduce a new prompting approach, analogical prompting, designed to automatically guide the reasoning process of large language models.
  arXiv  Detail & Related papers  (2023-10-03T00:57:26Z)
• Explainability for Large Language Models: A Survey [59.67574757137078]
  Large language models (LLMs) have demonstrated impressive capabilities in natural language processing. This paper introduces a taxonomy of explainability techniques and provides a structured overview of methods for explaining Transformer-based language models.
  arXiv  Detail & Related papers  (2023-09-02T22:14:26Z)
• Relational Concept Bottleneck Models [13.311396882130033]
  Concept Bottleneck Models (CBMs) are not designed to solve problems. R-CBMs are capable of both representing standard CBMs and relational GNNs. In particular, we show that R-CBMs support the generation of concept-based explanations.
  arXiv  Detail & Related papers  (2023-08-23T08:25:33Z)
• Incorporating Domain Knowledge in Deep Neural Networks for Discrete Choice Models [0.5801044612920815]
  This paper proposes a framework that expands the potential of data-driven approaches for DCM. It includes pseudo data samples that represent required relationships and a loss function that measures their fulfillment. A case study demonstrates the potential of this framework for discrete choice analysis.
  arXiv  Detail & Related papers  (2023-05-30T12:53:55Z)
• Scaling Pre-trained Language Models to Deeper via Parameter-efficient Architecture [68.13678918660872]
  We design a more capable parameter-sharing architecture based on matrix product operator (MPO) MPO decomposition can reorganize and factorize the information of a parameter matrix into two parts. Our architecture shares the central tensor across all layers for reducing the model size.
  arXiv  Detail & Related papers  (2023-03-27T02:34:09Z)
• When to Update Your Model: Constrained Model-based Reinforcement Learning [50.74369835934703]
  We propose a novel and general theoretical scheme for a non-decreasing performance guarantee of model-based RL (MBRL) Our follow-up derived bounds reveal the relationship between model shifts and performance improvement. A further example demonstrates that learning models from a dynamically-varying number of explorations benefit the eventual returns.
  arXiv  Detail & Related papers  (2022-10-15T17:57:43Z)
• Learning to Reason With Relational Abstractions [65.89553417442049]
  We study how to build stronger reasoning capability in language models using the idea of relational abstractions. We find that models that are supplied with such sequences as prompts can solve tasks with a significantly higher accuracy.
  arXiv  Detail & Related papers  (2022-10-06T00:27:50Z)
• Structured learning of rigid-body dynamics: A survey and unified view from a robotics perspective [5.597839822252915]
  We study supervised regression models that combine rigid-body mechanics with data-driven modelling techniques. We provide a unified view on the combination of data-driven regression models, such as neural networks and Gaussian processes, with analytical model priors.
  arXiv  Detail & Related papers  (2020-12-11T11:26:48Z)
• Text Modular Networks: Learning to Decompose Tasks in the Language of Existing Models [61.480085460269514]
  We propose a framework for building interpretable systems that learn to solve complex tasks by decomposing them into simpler ones solvable by existing models. We use this framework to build ModularQA, a system that can answer multi-hop reasoning questions by decomposing them into sub-questions answerable by a neural factoid single-span QA model and a symbolic calculator.
  arXiv  Detail & Related papers  (2020-09-01T23:45:42Z)

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.