Medical O1 Reasoning SFT: Distilled Dataset for Medical Problem-Solving

A distilled dataset from Deepseek-R1 based on medical verifiable problems, released on 2025/02/22. The dataset is designed for instruction tuning and Supervised Fine-Tuning (SFT) of language models. It was created by author vBasura and last updated on 2026-05-23.

Use Cases

• Supervised Fine-Tuning (SFT) of language models based on medical verifiable problems.
• Initializing models with reasoning chains from Deepseek-R1 as mentioned in the description.
• Training models on a mix of medical and general instruction data using the separate 'mix' file.

Strengths

• Dataset is specifically built on medical verifiable problems, suggesting a focus on factual accuracy.
• Includes reasoning chains distilled from the Deepseek-R1 model, providing structured problem-solving steps.

Limitations

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

Provenance

Source

vBasura on Hugging Face

Collection Method

Distilled from the Deepseek-R1 model based on medical verifiable problems.

Freshness

Last updated 2026-05-23 22:35:45; freshness should be verified.