README.md

metadata

pretty_name: Chomsky Corpus
license: other
language:
  - en
task_categories:
  - text-generation

Chomsky Corpus

Summary

A cleaned text dataset scraped from selected public sections of chomsky.info.

Source URLs

• https://chomsky.info/articles/
• https://chomsky.info/interviews/
• https://chomsky.info/letters/
• https://chomsky.info/talks/
• https://chomsky.info/debates/

Schema

• record_id: stable record identifier
• section: source section on the website
• article_title: title extracted from the listing page
• page_title: title extracted from the detail page when available
• article_date: normalized date when available
• content: cleaned page text
• content_length: number of characters in content
• content_hash: normalized content hash for deduplication
• index_url: listing page where the document was discovered

Cleaning

• whitespace normalization
• exact deduplication using normalized content hash
• invalid domain rejection
• missing title rejection
• missing content rejection
• short-content rejection

Quality Snapshot

• catalog_count: 1192
• raw_record_count: 1192
• validated_record_count: 1191
• rejected_record_count: 1
• deduped_record_count: 1186
• duplicate_count: 5
• min_content_length: 204
• median_content_length: 16075
• max_content_length: 2251683