# OpenBioLink/ThoughtSource

A central, open resource for data and tools related to chain-of-thought reasoning in large language models. Developed @ Samwald research group: https://samwald.info/

Repository: https://github.com/OpenBioLink/ThoughtSource
Canonical: https://ross.abutalabs.com/products/thoughtsource
Language: Jupyter Notebook
License: MIT
License Family: permissive
Topics: dataset, machine-learning, natural-language-processing, question-answering, reasoning
Last push: 2024-12-16T15:41:40+00:00

## Health v2 (maintenance only)
Score: 23/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 8, longevity 100
- inputs: {"age_days": 1566, "days_push": 625, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1015, forks 81 (observed 2026-08-28T04:03:14.050877+00:00)

## What it is
ThoughtSource is an open hub of datasets and tools for chain-of-thought reasoning in large language models, providing standardized dataloaders in Hugging Face Datasets format. It aggregates human- and AI-generated reasoning chains for question-answering datasets, including medical and commonsense QA.

## Use cases
- download chain-of-thought reasoning datasets for LLM research
- evaluate LLM reasoning on question-answering benchmarks
- get standardized CoT data in Hugging Face format
- study medical question answering with reasoning chains
- compare human vs AI-generated reasoning chains
- train models on chain-of-thought data

## When to choose
- you need curated chain-of-thought datasets in a unified format
- you research LLM reasoning or prompting
- you work on medical or commonsense QA with reasoning

## When to avoid
- you need a production inference or serving tool
- you want raw datasets without post-processing
- your task is unrelated to reasoning or QA

## Facets
- artifact type: dataset
- maturity: active
- function: machine-learning, nlp, data-science, prompt-engineering
- domain: large-language-models, machine-learning, healthcare
- platform: python
- tags: chain-of-thought, question-answering, reasoning, datasets, huggingface-datasets, llm-evaluation, medical-qa, natural-language-processing

## Member repositories
- OpenBioLink/ThoughtSource (main) score 23

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:14.050877+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T07:11:24.122090+00:00, confidence not recorded.
  - readme: https://github.com/OpenBioLink/ThoughtSource (fetched 2026-08-28T04:03:14.050877+00:00, sha d3924956c501)
- Data as of 2026-08-30T08:39:29.467469+00:00.