Do PhD-level LLMs Truly Grasp Elementary Addition? Probing Rule Learning vs. Memorization in Large Language Models (2504.05262v1)

Abstract: Despite high benchmark scores, LLMs often fail simple problem, raising a critical question: Do LLMs learn mathematical principles or merely memorize patterns? Rather than designing increasingly complex benchmarks like recent works, we investigate this using elementary two-integer addition ($0$ to $2^{64}$), probing two core properties: commutativity ($A+B=B+A$) and compositional generalization (via isomorphic symbolic mappings, e.g., $7 \rightarrow y$). While state-of-the-art LLMs achieve 73.8-99.8\% accuracy on numerical addition, performance collapses to $\leq$7.5\% under symbolic mapping, indicating failure to generalize learned rules. Non-monotonic performance scaling with digit count and frequent commutativity violations (over 1,700 cases of $A+B \neq B+A$) further support this. Explicitly providing addition rules degrades performance by 81.2\% on average, while self-explanation maintains baseline accuracy, suggesting LLM arithmetic processing is misaligned with human-defined principles. Our findings indicate current LLMs rely on memory pattern over genuine rule learning, highlighting architectural limitations and the need for new approaches to achieve true mathematical reasoning.