# Dissecting Physics Reasoning in Small Language Models: A Multi-Dimensional Analysis from an Educational Perspective

arXiv: https://arxiv.org/abs/2505.20707 · HTML: https://mlsanity.com/paper/2505.20707/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Spatial and embodied reasoning, one of 17 directions in LLM reasoning.
- Builds on 26 of its 52 references that are in the map.
- First posted on arXiv 27 May 2025.

## Placed in

- LLM reasoning / [Spatial and embodied reasoning](https://mlsanity.com/lr/spatial-and-embodied-reasoning/)

## Builds on (most cited first, 26 of 26)

- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [Measuring Massive Multitask Language Understanding](https://mlsanity.com/paper/2009.03300/) (2020)
- [DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning](https://mlsanity.com/paper/2501.12948/) (2025)
- [Qwen2.5 Technical Report](https://mlsanity.com/paper/2412.15115/) (2024)
- [Phi-3 Technical Report: A Highly Capable Language Model Locally on Your Phone](https://mlsanity.com/paper/2404.14219/) (2024)
- [Gemma 3 Technical Report](https://mlsanity.com/paper/2503.19786/) (2025)
- [A Survey on LLM-as-a-Judge](https://mlsanity.com/paper/2411.15594/) (2024)
- [OlympiadBench: A Challenging Benchmark for Promoting AGI with Olympiad-Level Bilingual Multimodal Scientific Problems](https://mlsanity.com/paper/2402.14008/) (2024)
- [Gemini: A Family of Highly Capable Multimodal Models](https://mlsanity.com/paper/2312.11805/) (2023)
- [Phi-4 Technical Report](https://mlsanity.com/paper/2412.08905/) (2024)
- [Measuring Faithfulness in Chain-of-Thought Reasoning](https://mlsanity.com/paper/2307.13702/) (2023)
- [GSM-Symbolic: Understanding the Limitations of Mathematical Reasoning in Large Language Models](https://mlsanity.com/paper/2410.05229/) (2024)
- [SciBench: Evaluating College-Level Scientific Problem-Solving Abilities of Large Language Models](https://mlsanity.com/paper/2307.10635/) (2023)
- [Judging the Judges: Evaluating Alignment and Vulnerabilities in LLMs-as-Judges](https://mlsanity.com/paper/2406.12624/) (2024)
- [Phi-4-reasoning Technical Report](https://mlsanity.com/paper/2504.21318/) (2025)
- [Have LLMs Advanced Enough? A Challenging Problem Solving Benchmark For Large Language Models](https://mlsanity.com/paper/2305.15074/) (2023)
- [J1: Incentivizing Thinking in LLM-as-a-Judge via Reinforcement Learning](https://mlsanity.com/paper/2505.10320/) (2025)
- [Evaluating Step-by-step Reasoning Traces: A Survey](https://mlsanity.com/paper/2502.12289/) (2025)
- [PhysReason: A Comprehensive Benchmark towards Physics-Based Reasoning](https://mlsanity.com/paper/2502.12054/) (2025)
- [Knowledge of cultural moral norms in large language models](https://mlsanity.com/paper/2306.01857/) (2023)
- [Towards Reasoning Ability of Small Language Models](https://mlsanity.com/paper/2502.11569/) (2025)
- [Localized Cultural Knowledge is Conserved and Controllable in Large Language Models](https://mlsanity.com/paper/2504.10191/) (2025)
- [Lost in Cultural Translation: Do LLMs Struggle with Math Across Cultural Contexts?](https://mlsanity.com/paper/2503.18018/) (2025)
- [WorldView-Bench: A Benchmark for Evaluating Global Cultural Perspectives in Large Language Models](https://mlsanity.com/paper/2505.09595/) (2025)
- [Small but Significant: On the Promise of Small Language Models for Accessible AIED](https://mlsanity.com/paper/2505.08588/) (2025)

## Abstract (arXiv)

Small Language Models (SLMs) offer privacy and efficiency for educational deployment, yet their utility depends on reliable multistep reasoning. Existing benchmarks often prioritize final answer accuracy, obscuring 'right answer, wrong procedure' failures that can reinforce student misconceptions. This work investigates SLM physics reasoning reliability, stage wise failure modes, and robustness under paired contextual variants. We introduce Physbench, comprising of 3,162 high school and AP level physics questions derived from OpenStax in a structured reference solution format with Bloom's Taxonomy annotations, plus 2,700 paired culturally contextualized variants. Using P-REFS, a stage wise evaluation rubric, we assess 10 SLMs across 58,000 responses. Results reveal substantial reliability gap: among final answer correct solutions, 75 to 98% contain at least one reasoning error. Failure modes shift with model capability; weaker models fail primarily at interpretation or modeling while stronger models often fail during execution. Paired contextual variations have minimal impact on top models but degrade the performance of mid-tier models. These findings demonstrate that safe educational AI requires evaluation paradigms that prioritize reasoning fidelity over final-answer correctness.