Evaluating the performance and robustness of LLMs in materials science Q&A and property predictions
We evaluate LLM performance and robustness for materials science Q&A and property prediction. Prompt sensitivity and mode collapse reveal reliability concerns, providing informed skepticism and guiding more cautious adoption in scientific research.
Authors
- Kangming Li (ORCID: https://orcid.org/0000-0003-4471-8527)
- Jason Hattrick‐Simpers (ORCID: https://orcid.org/0000-0003-2937-3188)
- Yao Fehlis
- Edward Kim (ORCID: https://orcid.org/0000-0002-9828-4624)
- Scott Ramsay
- Hongchen Wang
Institutions
- University of Toronto (CA)
Publication Details
- Journal
- Digital Discovery
- Published
- 2025-01-01
- DOI
- https://doi.org/10.1039/d5dd00090d
- Citations
- 4
- Primary Topic
- Machine Learning in Materials Science
- Type
- article
- Field-Weighted Citation Impact
- 2.07