Science Domain Evals
45 resources · page 2 of 2
Stanford HAI/RegLab: legal models hallucinate in 1 of 6+ queries
Creator not supplied; see original source.
Stanford HAI article on a RegLab study measuring hallucination rates of legal AI research tools such as Lexis+ AI and Westlaw AI.
TaxCalcBench (Column Tax)
Creator not supplied; see original source.
Column Tax post introducing TaxCalcBench, a benchmark of tax-return calculation cases with an analysis of where models fail.
Therapeutics Data Commons
Creator not supplied; see original source.
Therapeutics Data Commons: datasets, tasks and leaderboards for machine learning across drug discovery and development.
Vals.ai
Creator not supplied; see original source.
Vals AI benchmarks: independently run leaderboards on coding, finance, legal and science tasks with evaluation costs.
Vals AI's Latest Benchmark Finds Legal and General AI Now Outperform Lawyers in Legal Research (LawNext)
Creator not supplied; see original source.
LawNext article on Vals AI's legal research benchmark comparing legal AI tools and general LLMs with a lawyer baseline.
WeatherBench 2
Creator not supplied; see original source.
Google Research WeatherBench 2: benchmark, datasets and metrics for comparing data-driven global weather forecasting models.
WorldScore
Creator not supplied; see original source.
WorldScore project page: unified benchmark for 3D, 4D and video world generation on controllability, quality and dynamics.
Legal Technology Insider: Vals report missing the market leaders
Creator not supplied; see original source.
Legal Technology Insider article on the Vals AI legal research benchmark report, noting which major vendors did not take part.
MoleculeNet
Creator not supplied; see original source.
arXiv paper (2017) "MoleculeNet: A Benchmark for Molecular Machine Learning", introducing the MoleculeNet benchmark datasets for molecular property prediction.
FinanceBench: A New Benchmark for Financial Question Answering
Creator not supplied; see original source.
arXiv paper (2023) introducing FinanceBench, a benchmark for financial question answering over company filings.
FrontierMath paper
Creator not supplied; see original source.
arXiv paper (2024) introducing FrontierMath, a benchmark of unpublished advanced mathematics problems.
Medical VLM contamination audit
Creator not supplied; see original source.
arXiv paper (2026) auditing pretraining contamination of public medical vision-language benchmarks.
When Rubrics Fail
Creator not supplied; see original source.
arXiv PDF (2026) "When Rubrics Fail", a critique of rubric-based evaluation in medical LLM benchmarks.