PERC 2025 Abstract Detail Page
Previous Page | New Search | Browse All
| Abstract Title: | Reasoning models for problem solving in physics |
|---|---|
| Abstract Type: | Contributed Poster Presentation |
| Abstract: | Reasoning models are the new generation of Large Language Models (LLMs) capable of complex problem solving. The reliability of reasoning models in solving physics textbook end-of-chapter problems was tested by evaluating a sample of \(n = 5\) solutions generated by the state-of-the-art, OpenAI's \texttt{o3-mini} per each problem from 20 chapters of Halliday and Resnick (12th ed.) Vol. 1. In total, \(N = 408\) problems were given to the model and \(N \times n = 2,040\) generated solutions examined. The model successfully solved 94\% of the problems posed, excelling at easier topics in mechanics but struggling with the harder ones such as waves and thermodynamics. |
| Footnote: | This work is supported in part by the U.S. National Science Foundation grant 23000645. Opinions expressed are of the authors and not of the Foundation. |
| Session Time: | Poster Session B |
| Poster Number: | B-88 |
| Contributed Paper Record: | Contributed Paper Information |
| Contributed Paper Download: | Download Contributed Paper |
Author/Organizer Information | |
| Primary Contact: |
Amir Bralin Purdue University West Lafayette, IN 47907 Phone: 7657720524 |
| Co-Author(s) and Co-Presenter(s) |
N. Sanjay Rebello, Purdue University |
Contributed Poster | |
| Contributed Poster: | Download the Contributed Poster |
