Mixed Medical Reasoning Formatted for Instruction Tuning

A turn-based format dataset mixing FreedomIntelligence/Medical-R1-Distill-Data and its Chinese counterpart. It was created by ToastyPigeon and last updated on June 26, 2025. The data includes versions with and without explicit reasoning steps.

Use Cases

• Fine-tuning large language models for medical question-answering based on the described turn-based format.
• Training models to generate or skip explicit reasoning chains based on the '-thinking' and '-nothink' file variants.
• Developing multilingual medical AI assistants using the mixed English and Chinese source data.
• Studying the effect of reasoning step distillation on model performance in a clinical context.

Strengths

• Provides two distinct data variants ('-thinking' and '-nothink') for comparative training.
• Incorporates multilingual content from both English and Chinese medical reasoning sources.
• Last updated on June 26, 2025, suggesting recent maintenance.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

huggingface

Collection Method

Mixed and reformatted from FreedomIntelligence/Medical-R1-Distill-Data and FreedomIntelligence/Medical-R1-Distill-Data-Chinese.

Freshness

Last updated 2025-06-26 21:09:31