15,000 instruction-response pairs for training large language models on clinical tasks. The dataset is derived from the Alpaca and ShareGPT instruction-tuning datasets and formatted in JSONL. It appears to focus on medical domains such as SOAP notes, ICD coding, drug safety, and radiology.
Use Cases
- Fine-tuning LLMs for generating SOAP-style clinical notes based on the description.
- Training models for medical code (ICD) assignment tasks based on the description.
- Developing AI assistants for drug safety information retrieval based on the description.
- Benchmarking LLM performance on radiology report generation or analysis based on the description.
Strengths
- Contains 15,000 instruction-response pairs, providing a substantial volume for training.
- Focuses on multiple concrete clinical domains as mentioned in the description (SOAP, ICD coding, drug safety, radiology).
Limitations
- Column-level documentation is absent; field semantics must be inferred after download.
- Row count is unknown, which may limit suitability assessment.
- Last update date is unknown; freshness unverified.
Provenance
- Source
- null
- Collection Method
- Derived from Alpaca and ShareGPT instruction-tuning datasets.
- Time Range
- null
- Freshness
- null
- Geography
- null