google-research/deduplicate-text-datasets

• highabout#1

  Add a concise 'About' description

  Why:

  COPY-PASTE FIX

  Copy

  Code and scripts for ExactSubstr deduplication of large text datasets (e.g., C4, LM1B) to improve language model training efficiency and reduce memorization. Includes Rust implementation and Python analysis tools.

  OpenDoneDismiss
• mediumreadme#2

  Clarify the specific deduplication method offered in the README's opening

  Why:

  CURRENT

  This repository contains code to deduplicate language model datasets as descrbed in the paper "Deduplicating Training Data Makes Language Models Better" by Katherine Lee, Daphne Ippolito, Andrew Nystrom, Chiyuan Zhang, Douglas Eck, Chris Callison-Burch and Nicholas Carlini. We release the ExactSubstr deduplication implementation (written in Rust) along with the scripts we used in the paper to perform ExactSubstr deduplication and inspect the results (written in Python).

  COPY-PASTE FIX

  Copy

  This repository provides the **ExactSubstr deduplication implementation (written in Rust)** and associated Python scripts, as described in the paper "Deduplicating Training Data Makes Language Models Better". It enables efficient deduplication of large language model datasets to improve training and reduce memorization. We also release document clusters from NearDup deduplication on C4, RealNews, LM1B, and Wiki-4B-en.

  OpenDoneDismiss

Category GEO backends resolved for this scan: google/gemini-2.5-flash, deepseek/deepseek-v4-flash

• CATEGORY QUERY

  Seeking a method to efficiently deduplicate massive text corpora for training large language models.

  you: #2
  AI recommended (in order):

  1. datasketch
  2. deduplicate-text-datasets ← you
  3. text_dedup
  4. Apache Spark
  5. dedupe
  Show full AI answer
• CATEGORY QUERY

  What are robust tools for identifying and removing duplicate content from very large text datasets?

  you: not recommended
  AI recommended (in order):

  1. Apache Spark
  2. Spark's MLlib
  3. MinHashLSH
  4. Dedupe.io (dedupeio/dedupe)
  5. Datasketch (ekzhu/datasketch)
  6. Dask (dask/dask)
  7. PySpark
  8. Elasticsearch (elastic/elasticsearch)
  9. Python
  10. fuzzywuzzy (seatgeek/fuzzywuzzy)
  11. difflib
  12. OpenRefine (OpenRefine/OpenRefine)

  AI recommended 12 alternatives but never named google-research/deduplicate-text-datasets. This is the gap to close.

  Show full AI answer

• Metadata completeness

  fail

  Suggestion:

• README presence

  pass

• Compared to common alternatives in this category, what is the core differentiator of google-research/deduplicate-text-datasets?

  pass
  ⚠AI did not name google-research/deduplicate-text-datasets — likely talking about a different project

  AI answers can be confidently wrong. Read for accuracy: does it match your actual tech stack, audience, and differentiator?

• If a team adopts google-research/deduplicate-text-datasets in production, what risks or prerequisites should they evaluate first?

  pass
  ✓AI named google-research/deduplicate-text-datasets explicitly

  AI answers can be confidently wrong. Read for accuracy: does it match your actual tech stack, audience, and differentiator?

• In one sentence, what problem does the repo google-research/deduplicate-text-datasets solve, and who is the primary audience?

  pass
  ⚠AI did not name google-research/deduplicate-text-datasets — likely talking about a different project

  AI answers can be confidently wrong. Read for accuracy: does it match your actual tech stack, audience, and differentiator?