StreetMath: Study of LLMs' Approximation Behaviors

• URL: http://arxiv.org/abs/2510.25776v1
• Date: Mon, 27 Oct 2025 05:16:00 GMT
• Title: StreetMath: Study of LLMs' Approximation Behaviors
• Abstract summary: We introduce StreetMath, a benchmark designed to evaluate models' approximation abilities under real-world approximation scenarios.<n>Our analysis reveals that LLMs generally attempt to compute exact values or external tools even in tasks that call for approximation.<n>We argue that LLMs do not exhibit cognitive miserliness in the same way humans do in street math settings.
• Score: 1.4119508208285607
• License: http://creativecommons.org/licenses/by-nc-nd/4.0/
• Abstract: There is a substantial body of literature examining the mathematical reasoning capabilities of large language models (LLMs), particularly their performance on precise arithmetic operations in autoregressive architectures. However, their ability to perform approximate reasoning in informal, fast-paced mathematical operations has received far less attention, especially among non-autoregressive decoder models. Our work addresses this gap by introducing StreetMath, a benchmark designed to evaluate models' approximation abilities under real-world approximation scenarios. We conduct extensive evaluations across different LLM architectures: Qwen3-4B-Instruct-2507, Qwen3-4B-Thinking-2507, Dream-v0-Instruct-7B, Falcon-Mamba-7B-Instruct, and Mamba-GPT-3B. Furthermore, we apply mechanistic interpretability techniques to probe their internal computational states. Our analysis reveals that LLMs generally attempt to compute exact values or invoke external tools even in tasks that call for approximation. Moreover, while models sometimes reach the correct answer in early layers or steps, they still consume more tokens when solving approximation tasks. Additional experiments indicate that exact and approximate arithmetic operations rely on largely separate neural components. Drawing upon research on cognitive psychology, we argue that LLMs do not exhibit cognitive miserliness in the same way humans do in street math settings. We open source our work https://github.com/ctseng777/StreetMath

Related papers

• Evaluating SageMath-Augmented LLM Agents for Computational and Experimental Mathematics [1.166361205377345]
  We propose a ReAct-style agentic setup that combines LLM reasoning with verifiable feedback from SageMath.<n>We evaluate this agentic setup across frontier models for solving research-level mathematical problems from the RealMath benchmark.
  arXiv  Detail & Related papers  (2026-07-07T21:29:59Z)
• Automatically Generating Hard Math Problems from Hypothesis-Driven Error Analysis [16.008582390875656]
  We propose a new benchmark generation pipeline that uses AI-generated hypotheses to identify the specific math concepts and skills that LLMs struggle with.<n>Our pipeline is highly adaptable and can be applied beyond math to explore a wide range of LLM capabilities.
  arXiv  Detail & Related papers  (2026-04-06T03:27:48Z)
• AI Agents as Universal Task Solvers [94.49762121230042]
  We show that the optimal speed-up that a universal solver can achieve using past data is tightly related to their algorithmic information.<n>We argue that the key quantity to optimize when scaling reasoning models is time, whose critical role in learning has so far only been indirectly considered.
  arXiv  Detail & Related papers  (2025-10-14T02:17:54Z)
• ASyMOB: Algebraic Symbolic Mathematical Operations Benchmark [0.0]
  Large language models (LLMs) are rapidly approaching the level of proficiency in university-level symbolic mathematics.<n>We introduce ASyMOB, a novel assessment framework focused exclusively on symbolic manipulation.
  arXiv  Detail & Related papers  (2025-05-28T23:11:14Z)
• Open Eyes, Then Reason: Fine-grained Visual Mathematical Understanding in MLLMs [62.875934732547435]
  Current large language models (MLLMs) often underperform on mathematical problem-solving tasks that require fine-grained visual understanding.<n>In this paper, we evaluate the visual grounding capabilities of state-of-the-art MLLMs and reveal a significant negative correlation between visual grounding accuracy and problem-solving performance.<n>We propose a novel approach, SVE-Math, featuring a geometric-grounded vision encoder and a feature router that dynamically adjusts the contribution of hierarchical visual feature maps.
  arXiv  Detail & Related papers  (2025-01-11T04:08:44Z)
• Do NOT Think That Much for 2+3=? On the Overthinking of o1-Like LLMs [76.43407125275202]
  o1-like models can emulate human-like long-time thinking during inference.<n>This paper presents the first comprehensive study on the prevalent issue of overthinking in these models.<n>We propose strategies to mitigate overthinking, streamlining reasoning processes without compromising accuracy.
  arXiv  Detail & Related papers  (2024-12-30T18:55:12Z)
• Unraveling Arithmetic in Large Language Models: The Role of Algebraic Structures [2.8311048083168657]
  Large language models (LLMs) have demonstrated remarkable mathematical capabilities, largely driven by chain-of-thought (CoT) prompting.<n>We propose that LLMs learn arithmetic by capturing algebraic structures, such as commutativity and identity properties.
  arXiv  Detail & Related papers  (2024-11-25T10:23:11Z)
• Executing Arithmetic: Fine-Tuning Large Language Models as Turing Machines [7.695524275630717]
  Large Language Models (LLMs) have demonstrated remarkable capabilities across a wide range of natural language processing and reasoning tasks. We propose a Composable Arithmetic Execution Framework (CAEF) that enables LLMs to learn to execute step-by-step computations by emulating Turing Machines. In our evaluation, CAEF achieves nearly 100% accuracy across seven common mathematical operations on the LLaMA 3.1-8B model.
  arXiv  Detail & Related papers  (2024-10-10T13:23:49Z)
• AlphaMath Almost Zero: Process Supervision without Process [6.318873143509028]
  We propose an innovative framework, AlphaMath, that bypasses the need for process annotations by leveraging Monte Carlo Tree Search (MCTS) This framework focuses on unleashing the potential of a well-pretrained LLM to autonomously enhance its mathematical reasoning. The experimental results on both in-domain and out-of-domain datasets demonstrate that even without GPT-4 or human-annotated process supervision, our AlphaMath framework achieves comparable or superior results to previous state-of-the-art methods.
  arXiv  Detail & Related papers  (2024-05-06T15:20:30Z)
• GSM-Plus: A Comprehensive Benchmark for Evaluating the Robustness of LLMs as Mathematical Problem Solvers [68.77382332826167]
  Large language models (LLMs) have achieved impressive performance across various mathematical reasoning benchmarks. One essential and frequently occurring evidence is that when the math questions are slightly changed, LLMs can behave incorrectly. This motivates us to evaluate the robustness of LLMs' math reasoning capability by testing a wide range of question variations.
  arXiv  Detail & Related papers  (2024-02-29T15:26:14Z)
• InternLM-Math: Open Math Large Language Models Toward Verifiable Reasoning [98.53491178426492]
  We open-source our math reasoning LLMs InternLM-Math which is continue pre-trained from InternLM2. We unify chain-of-thought reasoning, reward modeling, formal reasoning, data augmentation, and code interpreter in a unified seq2seq format. Our pre-trained model achieves 30.3 on the MiniF2F test set without fine-tuning.
  arXiv  Detail & Related papers  (2024-02-09T11:22:08Z)
• Evaluating LLMs' Mathematical and Coding Competency through Ontology-guided Interventions [47.83142414018448]
  We focus on two popular reasoning tasks: arithmetic reasoning and code generation. We introduce (i) a general ontology of perturbations for math and coding questions, (ii) a semi-automatic method to apply these perturbations, and (iii) two datasets. We show a significant performance drop across all the models against perturbed questions.
  arXiv  Detail & Related papers  (2024-01-17T18:13:07Z)
• MARIO: MAth Reasoning with code Interpreter Output -- A Reproducible Pipeline [12.186691561822256]
  We postulate that the inherent nature of large language models (LLMs) presents challenges in modeling mathematical reasoning. This paper introduces a novel math dataset, enhanced with a capability to utilize a Python code interpreter. We propose a tentative, easily replicable protocol for the fine-tuning of math-specific LLMs.
  arXiv  Detail & Related papers  (2024-01-16T08:08:01Z)
• Evaluating Language Models for Mathematics through Interactions [116.67206980096513]
  We introduce CheckMate, a prototype platform for humans to interact with and evaluate large language models (LLMs) We conduct a study with CheckMate to evaluate three language models (InstructGPT, ChatGPT, and GPT-4) as assistants in proving undergraduate-level mathematics. We derive a taxonomy of human behaviours and uncover that despite a generally positive correlation, there are notable instances of divergence between correctness and perceived helpfulness.
  arXiv  Detail & Related papers  (2023-06-02T17:12:25Z)

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.