Dataset · UCSC-VLAA
Medical reasoning for LLMs
UCSC-VLAA/MedReason
A large set of medical tasks with step-by-step reasoning for explainable LLM answers.
What's inside
Contains text medical tasks and step-by-step reasoning built using structured knowledge and knowledge graphs. According to metadata, the volume ranges from 10k to 100k examples. The dataset is designed for training and verifying factual and explainable medical reasoning in LLMs.
What you can do with it
- Fine-tune a medical reasoning LLM
- Evaluate answer explainability in medical QA
- Collect SFT data with chain-of-thought reasoning
- Tasks
- Question answering
- Size
- 10K–100K rows
- Downloads this month
- 5.2K
- Likes
- 90
Models trained on it
- mradermacher/Sanad-1.0-i1-GGUF 2K
- mradermacher/Qwen3-8B-MedReasonPath-i1-GGUF 664
- mradermacher/Sanad-1.0-GGUF 421
- mradermacher/Sculptor-Qwen3_Med-Reasoning-GGUF 331
- mradermacher/Qwen3-8B-MedReasonPath-GGUF 282
- Melikshah/gpt-oss-20b-clinical-cot-ggufText generation 235
- Rumiii/Qwen3-8B-MedReasonPathText generation 78
- prithivMLmods/Sculptor-Qwen3_Med-ReasoningText generation 34
How to load it
from datasets import load_dataset
ds = load_dataset("UCSC-VLAA/MedReason")