# google-deepmind/rc-data

Question answering dataset featured in "Teaching Machines to Read and Comprehend

Repository: https://github.com/google-deepmind/rc-data
Canonical: https://ross.abutalabs.com/products/rc-data
Language: Python
License: Apache-2.0
License Family: permissive
Archived: true
Last push: 2017-04-26T15:05:19+00:00

## Health v2 (maintenance only)
Score: 10/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 3992, "days_push": 3416, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, archived
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1296, forks 239 (observed 2026-08-28T04:04:16.921438+00:00)

## What it is
A question answering corpus from DeepMind accompanying the paper 'Teaching Machines to Read and Comprehend' (Hermann et al., NIPS 2015). It includes a Python script to generate question/answer pairs from CNN and Daily Mail news articles downloaded via the Wayback Machine.

## Use cases
- train reading comprehension models on CNN/Daily Mail QA pairs
- benchmark machine reading comprehension algorithms
- generate question answering datasets from news articles
- evaluate abstractive and extractive summarization baselines
- download processed CNN/Daily Mail QA data for NLP research

## When to choose
- you need the original CNN/Daily Mail reading comprehension dataset from the NIPS 2015 paper
- you want to reproduce or cite Hermann et al.'s QA corpus in research
- you need a large-scale cloze-style question answering benchmark

## When to avoid
- you need a modern, actively maintained QA dataset
- you cannot work with Python 2.7-era scripts and Wayback Machine downloads
- you want conversational or multi-hop QA rather than cloze-style entity questions

## Facets
- artifact type: dataset
- maturity: maintenance
- function: nlp, machine-learning, data-generation
- domain: machine-learning, deep-learning
- platform: python, cli
- tags: question-answering, reading-comprehension, cnn-dailymail, benchmark-dataset, nips-2015, natural-language-processing, linux

## Member repositories
- google-deepmind/rc-data (main) score 10

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:16.921438+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 2026-08-30T04:53:57.480683+00:00, confidence not recorded.
  - readme: https://github.com/google-deepmind/rc-data (fetched 2026-08-28T04:04:16.921438+00:00, sha 6ebf26d9ffd7)
- Data as of 2026-08-30T08:39:29.467469+00:00.