FDARxBench: Benchmarking Regulatory and Clinical Reasoning on FDA Generic Drug Assessment

1. 1Curate a dataset of FDA drug label documents
2. 2Develop a benchmark to evaluate document-grounded question-answering models
3. 3Collaborate with regulatory assessors to ensure the benchmark is relevant and effective
4. 4Use FDARxBench to evaluate and improve the performance of language models in regulatory and clinical reasoning

Data scientists and AI engineers on a team can benefit from FDARxBench to evaluate and improve the performance of language models in regulatory and clinical reasoning, while regulatory assessors can use it to develop more accurate question-answering systems

💡 FDARxBench provides a real-world benchmark for evaluating the performance of language models in regulatory and clinical reasoning

📊 Introducing FDARxBench: a benchmark for evaluating document-grounded QA in generic drug assessment