Language Models as Science Tutors
Abstract
NLP has recently made exciting progress toward training language models (LMs) with strong scientific problem-solving skills. However, model development has not focused on real-life use-cases of LMs for science, including applications in education that require processing long scientific documents. To address this, we introduce TutorEval and TutorChat. TutorEval is a diverse question-answering benchmark consisting of questions about long chapters from STEM textbooks, written by experts. TutorEval helps measure real-life usability of LMs as scientific assistants, and it is the first benchmark combining long contexts, free-form generation, and multi-disciplinary scientific knowledge. Moreover, we show that fine-tuning base models with existing dialogue datasets leads to poor performance on TutorEval. Therefore, we create TutorChat, a dataset of 80,000 long synthetic dialogues about textbooks. We use TutorChat to fine-tune Llemma models with 7B and 34B parameters. These LM tutors specialized in math have a 32K-token context window, and they excel at TutorEval while performing strongly on GSM8K and MATH. Our datasets build on open-source materials, and we release our models, data, and evaluations.
Cite
@article{arxiv.2402.11111,
title = {Language Models as Science Tutors},
author = {Alexis Chevalier and Jiayi Geng and Alexander Wettig and Howard Chen and Sebastian Mizera and Toni Annala and Max Jameson Aragon and Arturo Rodríguez Fanlo and Simon Frieder and Simon Machado and Akshara Prabhakar and Ellie Thieu and Jiachen T. Wang and Zirui Wang and Xindi Wu and Mengzhou Xia and Wenhan Xia and Jiatong Yu and Jun-Jie Zhu and Zhiyong Jason Ren and Sanjeev Arora and Danqi Chen},
journal= {arXiv preprint arXiv:2402.11111},
year = {2024}
}
Comments
8 pages without bibliography and appendix, 26 pages total