Fine-Tuning Small Reasoning Models for Quantum Field Theory

• URL: http://arxiv.org/abs/2604.18936v1
• Date: Tue, 21 Apr 2026 00:21:05 GMT
• Title: Fine-Tuning Small Reasoning Models for Quantum Field Theory
• Abstract summary: We perform the first academic fine-tuning study of small reasoning models dedicated specifically to theoretical physics.<n>We generated over 2,500 synthetic problems alongside a curated collection of human-adapted problems sourced from arXiv and standard pedagogical resources.<n>We perform an extensive analysis of model chains-of-though before and after fine-tuning, to understand how reasoning errors evolve during RL and SFT.
• Score: 12.573129157671831
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Despite the growing application of Large Language Models (LLMs) to theoretical physics, there is little academic exploration into how domain-specific physics reasoning ability develops while training these models. To investigate this, we perform the first academic fine-tuning study of small (7B-parameter) reasoning models dedicated specifically to theoretical physics. Because open-source verifiable training data required to train such capabilities is scarce, we developed a robust data generation pipeline that can both create synthetic problems and make existing human-authored problems suitable for model training. Selecting Quantum Field Theory (QFT) as our primary domain, we generated over 2,500 synthetic problems alongside a curated collection of human-adapted problems sourced from arXiv and standard pedagogical resources. We conduct both Reinforcement Learning (RL) and Supervised Fine-Tuning (SFT) experiments, benchmarking performance gains as well as generalization to other physics domains. We perform an extensive analysis of model chains-of-though before and after fine-tuning, to understand how reasoning errors evolve during RL and SFT. Finally, we publicly release our data pipeline, verifiable QFT training data, and $\sim$200M tokens of QFT reasoning traces.

Related papers

• ArXiv-to-Model: A Practical Study of Scientific LM Training [0.16921396880325776]
  We present a case study of training a scientific language model directly from raw arXiv sources.<n>We analyze training stability, scaling behavior, data yield losses, and infrastructure bottlenecks.<n>This work provides an engineering-grounded, transparent account of training a small scientific language model from scratch.
  arXiv  Detail & Related papers  (2026-02-19T11:47:30Z)
• PhysProver: Advancing Automatic Theorem Proving for Physics [17.074001092418793]
  This paper presents the first approach to enhance formal theorem proving in the physics domain.<n>It is composed of theorems sampled from PhysLean and data generated by a conjecture-based formal data generation pipeline.<n>In the training pipeline, we leverage DeepSeek-Prover-V2-7B, a strong open-source mathematical theorem prover, and apply Reinforcement Learning with Verifiable Rewards (RLVR) to train our model PhysProver.
  arXiv  Detail & Related papers  (2026-01-22T08:05:32Z)
• Quantum-Aware Generative AI for Materials Discovery: A Framework for Robust Exploration Beyond DFT Biases [0.0]
  We introduce a quantum-aware generative AI framework for materials discovery.<n>We implement a robust active learning loop that quantifies and targets the divergence between low- and high-fidelity predictions.<n>Our results demonstrate a 3-5x improvement in successfully identifying potentially stable candidates in high-divergence regions.
  arXiv  Detail & Related papers  (2025-12-13T11:17:21Z)
• Benchmarking Foundation Models with Retrieval-Augmented Generation in Olympic-Level Physics Problem Solving [56.119382216818195]
  Retrieval-augmented generation (RAG) with foundation models has achieved strong performance across diverse tasks.<n>But their capacity for expert-level reasoning-such as solving Olympiad-level physics problems-remains largely unexplored.<n>We introduce PhoPile, a high-quality multimodal dataset specifically designed for Olympiad-level physics.<n>Using PhoPile, we benchmark RAG-augmented foundation models, covering both large language models (LLMs) and large multimodal models (LMMs) with multiple retrievers.
  arXiv  Detail & Related papers  (2025-10-01T13:57:53Z)
• A Theoretical Perspective: How to Prevent Model Collapse in Self-consuming Training Loops [55.07063067759609]
  High-quality data is essential for training large generative models, yet the vast reservoir of real data available online has become nearly depleted.<n>Models increasingly generate their own data for further training, forming Self-consuming Training Loops (STLs)<n>Some models degrade or even collapse, while others successfully avoid these failures, leaving a significant gap in theoretical understanding.
  arXiv  Detail & Related papers  (2025-02-26T06:18:13Z)
• Unleashing LLM Reasoning Capability via Scalable Question Synthesis from Scratch [54.12139707822201]
  We propose ScaleQuest, a novel, scalable, and cost-effective data synthesis method.<n>By generating diverse questions from scratch, we produce a dataset of 1 million problem-solution pairs.<n>Our experiments demonstrate that models trained on our data outperform existing open-source datasets.
  arXiv  Detail & Related papers  (2024-10-24T12:42:04Z)
• Physics-Informed Weakly Supervised Learning for Interatomic Potentials [17.165117198519248]
  We introduce a physics-informed, weakly supervised approach for training machine-learned interatomic potentials (MLIPs)<n>We demonstrate reduced energy and force errors -- often lower by a factor of two -- for various baseline models and benchmark data sets.<n>Our approach improves the fine-tuning of foundation models on sparse, highly accurate ab initio data.
  arXiv  Detail & Related papers  (2024-07-23T12:49:04Z)
• Discovering Interpretable Physical Models using Symbolic Regression and Discrete Exterior Calculus [55.2480439325792]
  We propose a framework that combines Symbolic Regression (SR) and Discrete Exterior Calculus (DEC) for the automated discovery of physical models. DEC provides building blocks for the discrete analogue of field theories, which are beyond the state-of-the-art applications of SR to physical problems. We prove the effectiveness of our methodology by re-discovering three models of Continuum Physics from synthetic experimental data.
  arXiv  Detail & Related papers  (2023-10-10T13:23:05Z)
• Human Trajectory Prediction via Neural Social Physics [63.62824628085961]
  Trajectory prediction has been widely pursued in many fields, and many model-based and model-free methods have been explored. We propose a new method combining both methodologies based on a new Neural Differential Equation model. Our new model (Neural Social Physics or NSP) is a deep neural network within which we use an explicit physics model with learnable parameters.
  arXiv  Detail & Related papers  (2022-07-21T12:11:18Z)
• Training a First-Order Theorem Prover from Synthetic Data [50.23600875138756]
  A major challenge in applying machine learning to automated theorem proving is the scarcity of training data. We propose an approach that relies on training purely with synthetically generated theorems, without any human data aside from axioms. Our neural prover outperforms the state-of-the-art E-prover on this synthetic data in both time and search steps.
  arXiv  Detail & Related papers  (2021-03-05T17:01:34Z)

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.