MoveBench evaluates 15 LLMs on sports and health coaching
Original title: MoveBench: A Rubric-Based Benchmark for Evaluating Large Language Models in Sports and Health Coaching
Models were scored on 809 questions and 5,012 rubric criteria covering safety, validity, and personalization.
MoveBench is a rubric-based benchmark for large language models used in sports and health coaching, covering multiple-choice and open-ended tasks grounded in user scenarios. Rubrics emphasize safety compliance, scientific validity, and training logic. The authors evaluated 15 LLMs on 809 questions with 5,012 rubric-level criteria and report wide variation in safety awareness, program design, and personalization. Resources are published at the project's GitHub repository.
Unconfirmed: Performance gaps are reported by the benchmark authors; individual model scores are not given here.
- The set has 809 questions and 5,012 rubric-level criteria.
- Fifteen diverse LLMs were evaluated.
- Tasks include MCQs and open-ended coaching questions.
- Authors flag gaps in safety awareness and personalization.
From the original article
1 min read at huggingface.co
Despite rapid advances in large language models (LLMs), their deployment in the sports and health domain remains high-risk due to the absence of domain-specific evaluation benchmarks. Existing general-purpose benchmarks (e. g.
By Yitong Li, Jiang Bian, Jingling Wang · huggingface.co
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.