MatSciBench: Materials Science Questions for LLM Evaluation

A benchmark dataset for materials science question answering created by JunkaiZ. The dataset contains questions spanning domains such as polymer science, crystallography, and thermodynamics. It was last updated on 2026-05-31.

Use Cases

• Benchmarking LLM performance based on questions across materials science subfields
• Training domain-specific QA models based on the described question-answer pairs
• Analyzing model reasoning gaps based on the described coverage of topics like thermodynamics and crystallography

Strengths

• Covers multiple domains within materials science, including polymer science, crystallography, and thermodynamics
• Designed specifically for benchmarking large language models

Limitations

• Column-level documentation is absent; field semantics must be inferred after download
• Row count is unknown, which may limit suitability assessment

Provenance

Source

JunkaiZ via Hugging Face

Collection Method

Created for the paper 'MatSciBench: Benchmarking the Reasoning Ability of Large Language Models in Materials Science'

Freshness

Last updated 2026-05-31 08:51:56