NLP Lineage

31 December 2020 · Language modelling

The Pile

The Pile: An 800GB Dataset of Diverse Text

L. Gao, S. Biderman, S. Black, et al. (12 authors)

Date
31 December 2020
What the date marks
arXiv submission
Kind
Dataset
Organisation
EleutherAI
Access
mixed, openly distributed
Threads
In the guided tour
Chapter 5. Scale becomes a science
Learning phases
2. How LLMs are trained; 8. Training from scratch

A milestone: 5 later milestones descend directly from it, and its descendants reach 4 of the 10 threads. How levels are decided

Read the primary source (arxiv.org)

Full summary in progress. Until then, this page shows the milestone's record from the research data.

What it changed

Before it

Frontier training data was proprietary and undocumented.

The problem

Outsiders could not reproduce large-model training without a comparable corpus.

What it introduced

825 GiB assembled from 22 documented sources, released openly.

Why it matters

It let independent groups train serious models, which is how the open ecosystem started.

What modern models inherit

Open, documented pretraining data as a community resource.

What came next

Pythia, GPT-NeoX and OPT trained on it; later corpora refined its mix.

Before and after

The milestones it descends from, and the milestones that descend from it. Work developed alongside it is not counted. Weak links are left out unless you include them.

Before: where it came from

The Pile has 19 documented ancestors across 5 chapters, reaching back to Recurrent networks in 1990. Showing its direct parents and the 10 turning points among them.

The Pile has 19 documented ancestors across 5 chapters, reaching back to Recurrent networks in 1990. Showing all of them.

The Pile has 28 recorded ancestors, including indirect ones, across 5 chapters, reaching back to Recurrent networks in 1990. Showing its direct parents and the 10 turning points among them.

The Pile has 28 recorded ancestors, including indirect ones, across 5 chapters, reaching back to Recurrent networks in 1990. Showing all of them.

  1. 2020The PileThis milestone.
  2. 2020GPT-3 Turning pointThe Pile was influenced by thisChapter 5. Scale becomes a science
  3. 2020Scaling laws Turning pointGPT-3 was influenced by thisChapter 5. Scale becomes a science
  4. 2019C4 Turning pointThe Pile was influenced by thisChapter 4. Pretrain once, use everywhere
  5. 2019Sparse Transformer Turning pointGPT-3 built directly on thisChapter 4. Pretrain once, use everywhere
  6. 2019GPT-2 Turning pointGPT-3 built directly on thisChapter 4. Pretrain once, use everywhere
  7. 2019WebTextGPT-2 was influenced by thisChapter 4. Pretrain once, use everywhere
  8. 2018GPT-1GPT-2 built directly on thisChapter 4. Pretrain once, use everywhere
  9. 2017Transformer Turning pointGPT-2 built directly on this
    Sparse Transformer built directly on this
    Scaling laws built directly on this    Chapter 3. Attention, then only attention
  10. 2016Layer normGPT-2 built directly on thisChapter 3. Attention, then only attention
  11. 2015Residual connectionsTransformer was influenced by thisChapter 3. Attention, then only attention
  12. 2015LM pretraining Turning pointGPT-1 was influenced by thisChapter 2. Words become vectors
  13. 2015BPE tokenizationGPT-2 built directly on thisChapter 2. Words become vectors
  14. 2015Luong attentionTransformer was influenced by thisChapter 3. Attention, then only attention
  15. 2014AdamTransformer was influenced by thisChapter 2. Words become vectors
  16. 2014Seq2seq Turning pointTransformer built directly on thisChapter 2. Words become vectors
  17. 2014Attention Turning pointTransformer built directly on thisChapter 3. Attention, then only attention
  18. 2014GRU / encoder-decoderAttention built directly on thisChapter 2. Words become vectors
  19. 1997LSTM Turning pointSeq2seq built directly on thisChapter 1. Language as probability
  20. 1990Recurrent networksLSTM was influenced by thisChapter 1. Language as probability
  1. 2020The PileThis milestone.
  2. 2020GPT-3 Turning pointThe Pile was influenced by thisChapter 5. Scale becomes a science
  3. 2020Scaling laws Turning pointGPT-3 was influenced by thisChapter 5. Scale becomes a science
  4. 2019C4 Turning pointThe Pile was influenced by thisChapter 4. Pretrain once, use everywhere
  5. 2019Sparse Transformer Turning pointGPT-3 built directly on thisChapter 4. Pretrain once, use everywhere
  6. 2019GPT-2 Turning pointGPT-3 built directly on thisChapter 4. Pretrain once, use everywhere
  7. 2019WebTextC4 was influenced by this (indirect influence)Chapter 4. Pretrain once, use everywhere
  8. 2019Transformer-XLSparse Transformer was influenced by this (indirect influence)Chapter 4. Pretrain once, use everywhere
  9. 2018GPT-1GPT-2 built directly on this
    WebText was influenced by this    Chapter 4. Pretrain once, use everywhere
  10. 2018ELMoGPT-1 was influenced by this (indirect influence)Chapter 4. Pretrain once, use everywhere
  11. 2018ULMFiTGPT-1 was influenced by this (indirect influence)Chapter 4. Pretrain once, use everywhere
  12. 2017CoVeELMo was influenced by thisChapter 3. Attention, then only attention
  13. 2017Transformer Turning pointGPT-2 built directly on this
    Sparse Transformer built directly on this
    Scaling laws built directly on this    Chapter 3. Attention, then only attention
  14. 2017ConvS2STransformer was influenced by this (indirect influence)Chapter 3. Attention, then only attention
  15. 2016Layer normGPT-2 built directly on thisChapter 3. Attention, then only attention
  16. 2015Residual connectionsTransformer was influenced by thisChapter 3. Attention, then only attention
  17. 2015LM pretraining Turning pointGPT-1 was influenced by thisChapter 2. Words become vectors
  18. 2015BPE tokenizationGPT-2 built directly on thisChapter 2. Words become vectors
  19. 2015Luong attentionTransformer was influenced by thisChapter 3. Attention, then only attention
  20. 2014AdamTransformer was influenced by thisChapter 2. Words become vectors
  21. 2014GloVeELMo built directly on thisChapter 2. Words become vectors
  22. 2014Seq2seq Turning pointTransformer built directly on thisChapter 2. Words become vectors
  23. 2014Attention Turning pointTransformer built directly on thisChapter 3. Attention, then only attention
  24. 2014GRU / encoder-decoderAttention built directly on thisChapter 2. Words become vectors
  25. 2013word2vecELMo was influenced by thisChapter 2. Words become vectors
  26. 2011NLP from scratchELMo was influenced by thisChapter 1. Language as probability
  27. 2003Neural LMNLP from scratch was influenced by this
    word2vec was influenced by this    Chapter 1. Language as probability
  28. 1997LSTM Turning pointSeq2seq built directly on this
    ELMo built directly on this    Chapter 1. Language as probability
  29. 1990Recurrent networksLSTM was influenced by thisChapter 1. Language as probability

After: what it led to

The Pile has 9 documented descendants across 2 chapters, reaching forward to FineWeb in 2024. Showing its direct descendants.

The Pile has 9 documented descendants across 2 chapters, reaching forward to FineWeb in 2024. Showing all of them.

The Pile has 10 recorded descendants, including indirect ones, across 2 chapters, reaching forward to FineWeb in 2024. Showing its direct descendants.

The Pile has 10 recorded descendants, including indirect ones, across 2 chapters, reaching forward to FineWeb in 2024. Showing all of them.

  1. 2020The PileThis milestone.
  2. 2022OPT-175Bbuilt directly on The PileChapter 8. The weights get out
  3. 2022BLOOMwas influenced by The PileChapter 8. The weights get out
  4. 2022LLM.int8()was influenced by OPT-175BChapter 9. Making it run anywhere
  5. 2022GPTQwas influenced by LLM.int8()Chapter 9. Making it run anywhere
  6. 2023Pythiabuilt directly on The PileChapter 8. The weights get out
  7. 2023QLoRAbuilt directly on LLM.int8()Chapter 9. Making it run anywhere
  8. 2024Dolmawas influenced by The PileChapter 8. The weights get out
  9. 2024OLMobuilt directly on Dolma
    was influenced by Pythia    Chapter 8. The weights get out
  10. 2024FineWebwas influenced by The PileChapter 8. The weights get out
  1. 2020The PileThis milestone.
  2. 2022OPT-175Bbuilt directly on The PileChapter 8. The weights get out
  3. 2022BLOOMwas influenced by The PileChapter 8. The weights get out
  4. 2022LLM.int8()was influenced by OPT-175BChapter 9. Making it run anywhere
  5. 2022GPTQwas influenced by LLM.int8()Chapter 9. Making it run anywhere
  6. 2023llama.cppwas influenced by GPTQ (indirect influence)Chapter 9. Making it run anywhere
  7. 2023Pythiabuilt directly on The PileChapter 8. The weights get out
  8. 2023QLoRAbuilt directly on LLM.int8()Chapter 9. Making it run anywhere
  9. 2024Dolmawas influenced by The PileChapter 8. The weights get out
  10. 2024OLMobuilt directly on Dolma
    was influenced by Pythia    Chapter 8. The weights get out
  11. 2024FineWebwas influenced by The PileChapter 8. The weights get out

Direct links and their sources

Where it came from

Influenced by

What it led to

Built on directly by

Influenced