metadata
license: cc-by-sa-4.0
WHB Subset
A curated evaluation dataset for assessing large language models (LLMs) in women’s health–related tasks. The dataset consists of model stumps paired with expert-written justifications describing observed errors.
The dataset was developed in collaboration with women’s health experts, including clinicians, pharmacists, and researchers. The model stumps represent realistic prompts drawn from:
- Patient-facing scenarios
- Clinician decision-making contexts
- Evidence-based medical questions
Each prompt is paired with an expert justification explaining the identified issue.
Dataset Description
This subset is derived from the Women’s Health Benchmark (WHB) and contains 20 representative model stumps (expert prompts) with corresponding expert justifications.
Dataset Statistics
- Model stumps: 20
- Expert justifications: 20
- Language: English
- Generation date: November 20, 2025