Representation Robustness Under Executable Reasoning Constraints in Large Language Models for Mathematical Problem Solving
Summary
This paper investigates representation robustness in LLMs for mathematical problem solving by systematically varying surface representations of equivalent problems, finding substantial sensitivity and showing that code-augmented reasoning does not uniformly eliminate brittleness.
View Cached Full Text
Cached at: 07/24/26, 05:04 AM
# Representation Robustness Under Executable Reasoning Constraints in Large Language Models for Mathematical Problem Solving Source: [https://arxiv.org/abs/2607.20520](https://arxiv.org/abs/2607.20520) [View PDF](https://arxiv.org/pdf/2607.20520) > Abstract:Large language models \(LLMs\) are increasingly evaluated on mathematical problem solving, yet prior work often treats representationally equivalent formulations as interchangeable and conflates reasoning errors with interface failures\. This paper investigates representation robustness in LLM\-based mathematical problem solving by systematically varying surface representations of the same underlying problems, including story problems, word\-equations, symbolic equations, and isomorphic paraphrases\. Using a curated dataset of mathematically equivalent problems, we evaluate five contemporary LLMs under a direct answer generation condition\. We find substantial representational sensitivity: models frequently change correctness across equivalent formulations, with nontrivial flip rates across story, symbolic, and word\-equation variants\. We also observe systematic regressions under isomorphic reformulations, showing that even subtle paraphrase\-level changes can degrade performance despite preserved mathematical structure\. We then evaluate a code\-augmented condition in which models externalize reasoning as executable Python code that is run locally for validation\. This interface reveals strong latent reasoning capability in some models that perform poorly under direct prompting, but it does not uniformly improve robustness\. Instead, failures shift across interaction layers, from opaque reasoning errors to protocol violations and execution failures\. Even when executable reasoning succeeds, representation sensitivity often persists\. Overall, our results show that reasoning scaffolds do not eliminate representational brittleness, but expose new tradeoffs among correctness, reliability, latency, and cost\. We argue that representation should be treated as a first\-class interface design variable in LLM evaluation and deployment, especially for AI\-assisted problem\-solving systems\. ## Submission history From: Sagnik Nath \[[view email](https://arxiv.org/show-email/6227b209/2607.20520)\] **\[v1\]**Wed, 8 Jul 2026 17:21:28 UTC \(1,013 KB\)
Similar Articles
Mathematical Reasoning in Large Language Models: Benchmarks, Architectures, Evaluation, and Open Challenges
This survey synthesizes recent advancements in mathematical reasoning with large language models, covering benchmarks, architectures, training strategies, and evaluation protocols. It identifies key challenges such as reasoning faithfulness and benchmark biases.
Disentangling Mathematical Reasoning in LLMs: A Methodological Investigation of Internal Mechanisms
This paper investigates how large language models perform arithmetic operations by analyzing internal mechanisms through early decoding, revealing that proficient models exhibit a clear division of labor between attention and MLP modules in reasoning tasks.
Measuring Representation Robustness in Large Language Models for Geometry
Researchers introduce GeoRepEval, a framework to evaluate LLM robustness across equivalent geometric problem representations (Euclidean, coordinate, vector). Testing 11 LLMs on 158 geometry problems, they find accuracy gaps up to 14 percentage points based solely on representation choice, with vector formulations being a consistent failure point.
Reasoning, Code, or Both? How Large Language Models Handle Variations in Math Questions
This paper evaluates three approaches (pure chain-of-thought reasoning, single-shot code execution, and iterative code execution) on 1,000 GSM-Symbolic problems using Claude Haiku 4.5, finding that chain-of-thought is the most robust to perturbation, while code execution does not improve reasoning robustness on grade-school math problems.
TabularMath: Understanding Math Reasoning over Tables with Large Language Models
TabularMath introduces a benchmark and AutoT2T framework for evaluating LLMs' mathematical reasoning over tabular data, revealing that table complexity, data quality, and modality significantly impact model performance. The study addresses a gap in LLM evaluation by systematically assessing robustness to incomplete or inconsistent table information in real-world scenarios.