All datasets / sbussiso

sbussiso/SmolThinker-Synthetic-Reasoning

64 downloads in 30 days1 likesLast modified 11 Aug 2026View on Hugging Face

Permissive or share-alike

The declared licence (mit) is classified as permissive.

Declared licence: mit, from the tagEvidence: README.md at f25af2c
Licence files: No licence file at the top level of the repositoryEvidence: file list at f25af2c
Declared sources: 2
Access gated on Hugging Face: No

Each cell links to what was checked: a file, a file list or a Hub record.

Sources listed in the card metadata

mattwesney/General_Inquiry_Thinking-Chain-Of-ThoughtmitPermissiveEvidence: README.md at 033ff55
nvidia/OpenScienceReasoning-2cc-by-4.0PermissiveEvidence: README.md at 174b02c

Tasks

text-generation

Citation text from the card

**[mattwesney/General_Inquiry_Thinking-Chain-Of-Thought](https://huggingface.co/datasets/mattwesney/General_Inquiry_Thinking-Chain-Of-Thought)**
(MIT) supplies the 600 long-form rows. Please star the original, which has
6,000 examples against the 600 sampled here. Rows were filtered for broken
`<think>` tags and length, normalised to a consistent tag shape, and converted
from Alpaca to a messages array.

**[nvidia/OpenScienceReasoning-2](https://huggingface.co/datasets/nvidia/OpenScienceReasoning-2)**
(CC-BY-4.0) supplies the 600 science and humanities rows. Credit for the
questions and the DeepSeek-R1 reasoning traces belongs to NVIDIA and the
DeepSeek-R1 authors. From a 20,000-row sample the shortest traces were selected
(median ~750 tokens), trimmed of leading and trailing boilerplate but never of
reasoning steps, and bare-letter answers restated as the full chosen option.
Option/letter alignment was re-verified on every row.

The remaining 6,596 rows are original and released under MIT.

All datasets marked Permissive or share-alike | JSON

The same facts for your agent

A free MCP server at mcp.datalicense.uk/mcp. Connect it from an MCP client such as Claude or Cursor.

dataset_licence({ "id": "sbussiso/SmolThinker-Synthetic-Reasoning" })
signal: "permissive"
stated_licence: "mit"
sources: 2