Saved in:
Bibliographic Details
Main Author: Wang, Charles L.
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2509.23143
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914178122383360
author Wang, Charles L.
author_facet Wang, Charles L.
contents This paper presents MathBode, a dynamic diagnostic for mathematical reasoning in large language models (LLMs). Instead of one-shot accuracy, MathBode treats each parametric problem as a system: we drive a single parameter sinusoidally and fit first-harmonic responses of model outputs and exact solutions. This yields interpretable, frequency-resolved metrics -- gain (amplitude tracking) and phase (lag) -- that form Bode-style fingerprints. Across five closed-form families (linear solve, ratio/saturation, compound interest, 2x2 linear systems, similar triangles), the diagnostic surfaces systematic low-pass behavior and growing phase lag that accuracy alone obscures. We compare several models against a symbolic baseline that calibrates the instrument ($G \approx 1$, $ϕ\approx 0$). Results separate frontier from mid-tier models on dynamics, providing a compact, reproducible protocol that complements standard benchmarks with actionable measurements of reasoning fidelity and consistency. We open-source the dataset and code to enable further research and adoption.
format Preprint
id arxiv_https___arxiv_org_abs_2509_23143
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle MathBode: Measuring the Stability of LLM Reasoning using Frequency Response
Wang, Charles L.
Artificial Intelligence
Machine Learning
Systems and Control
This paper presents MathBode, a dynamic diagnostic for mathematical reasoning in large language models (LLMs). Instead of one-shot accuracy, MathBode treats each parametric problem as a system: we drive a single parameter sinusoidally and fit first-harmonic responses of model outputs and exact solutions. This yields interpretable, frequency-resolved metrics -- gain (amplitude tracking) and phase (lag) -- that form Bode-style fingerprints. Across five closed-form families (linear solve, ratio/saturation, compound interest, 2x2 linear systems, similar triangles), the diagnostic surfaces systematic low-pass behavior and growing phase lag that accuracy alone obscures. We compare several models against a symbolic baseline that calibrates the instrument ($G \approx 1$, $ϕ\approx 0$). Results separate frontier from mid-tier models on dynamics, providing a compact, reproducible protocol that complements standard benchmarks with actionable measurements of reasoning fidelity and consistency. We open-source the dataset and code to enable further research and adoption.
title MathBode: Measuring the Stability of LLM Reasoning using Frequency Response
topic Artificial Intelligence
Machine Learning
Systems and Control
url https://arxiv.org/abs/2509.23143