NLP Lineage

20 December 2022 · Scale and data

Self-Instruct

Self-Instruct: Aligning Language Models with Self-Generated Instructions

Y. Wang, Y. Kordi, S. Mishra, A. Liu, N. Smith, D. Khashabi, H. Hajishirzi

Date
20 December 2022
What the date marks
arXiv submission
Kind
Training method
Organisation
U. Washington / AI2
Access
Apache 2.0
Other dates
ACL 2023
Threads
In the guided tour
Chapter 6. Teaching models what people want
Learning phases
2. How LLMs are trained; 5. Fine-tuning

A milestone: 1 later milestone descends directly from it, and its descendants reach 2 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

Instruction tuning needed human-written instruction-response pairs.

The problem

Writing tens of thousands of examples by hand is slow and expensive.

What it introduced

Bootstrap from a small seed set: have the model write new instructions and answers, filter, and train on the result.

Why it matters

It made instruction data nearly free, which is how small open models caught up so quickly in 2023.

What modern models inherit

Synthetic instruction data as a normal ingredient.

What came next

Alpaca applied it to GPT-3.5 outputs; the synthetic data line runs on to phi and beyond.

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

Self-Instruct has 30 documented ancestors across 6 chapters, reaching back to Recurrent networks in 1990. Showing its direct parents and the 12 turning points among them.

Self-Instruct has 30 documented ancestors across 6 chapters, reaching back to Recurrent networks in 1990. Showing all of them.

Self-Instruct has 32 recorded ancestors, including indirect ones, across 6 chapters, reaching back to Recurrent networks in 1990. Showing its direct parents and the 12 turning points among them.

Self-Instruct has 32 recorded ancestors, including indirect ones, across 6 chapters, reaching back to Recurrent networks in 1990. Showing all of them.

  1. 2022Self-InstructThis milestone.
  2. 2021FLANSelf-Instruct was influenced by thisChapter 6. Teaching models what people want
  3. 2020GPT-3 Turning pointSelf-Instruct built directly on thisChapter 5. Scale becomes a science
  4. 2020Scaling laws Turning pointGPT-3 was influenced by thisChapter 5. Scale becomes a science
  5. 2019T5 Turning pointFLAN was influenced by thisChapter 4. Pretrain once, use everywhere
  6. 2019C4 Turning pointT5 was influenced by thisChapter 4. Pretrain once, use everywhere
  7. 2019Sparse Transformer Turning pointGPT-3 built directly on thisChapter 4. Pretrain once, use everywhere
  8. 2019GPT-2 Turning pointGPT-3 built directly on thisChapter 4. Pretrain once, use everywhere
  9. 2019WebTextGPT-2 was influenced by thisChapter 4. Pretrain once, use everywhere
  10. 2018BERT Turning pointT5 was influenced by thisChapter 4. Pretrain once, use everywhere
  11. 2018SentencePieceT5 built directly on thisChapter 4. Pretrain once, use everywhere
  12. 2018GPT-1GPT-2 built directly on thisChapter 4. Pretrain once, use everywhere
  13. 2018ELMoBERT was influenced by thisChapter 4. Pretrain once, use everywhere
  14. 2018ULMFiTBERT was influenced by thisChapter 4. Pretrain once, use everywhere
  15. 2017CoVeELMo was influenced by thisChapter 3. Attention, then only attention
  16. 2017Transformer Turning pointGPT-2 built directly on this
    Sparse Transformer built directly on this
    T5 built directly on this    Chapter 3. Attention, then only attention
  17. 2016Layer normGPT-2 built directly on thisChapter 3. Attention, then only attention
  18. 2015Residual connectionsTransformer was influenced by thisChapter 3. Attention, then only attention
  19. 2015LM pretraining Turning pointBERT was influenced by this
    GPT-1 was influenced by this    Chapter 2. Words become vectors
  20. 2015BPE tokenizationGPT-2 built directly on thisChapter 2. Words become vectors
  21. 2015Luong attentionTransformer was influenced by thisChapter 3. Attention, then only attention
  22. 2014AdamTransformer was influenced by thisChapter 2. Words become vectors
  23. 2014GloVeELMo built directly on thisChapter 2. Words become vectors
  24. 2014Seq2seq Turning pointTransformer built directly on thisChapter 2. Words become vectors
  25. 2014Attention Turning pointTransformer built directly on thisChapter 3. Attention, then only attention
  26. 2014GRU / encoder-decoderAttention built directly on thisChapter 2. Words become vectors
  27. 2013word2vecELMo was influenced by thisChapter 2. Words become vectors
  28. 2011NLP from scratchELMo was influenced by thisChapter 1. Language as probability
  29. 2003Neural LMNLP from scratch was influenced by this
    word2vec was influenced by this    Chapter 1. Language as probability
  30. 1997LSTM Turning pointSeq2seq built directly on this
    ELMo built directly on this    Chapter 1. Language as probability
  31. 1990Recurrent networksLSTM was influenced by thisChapter 1. Language as probability
  1. 2022Self-InstructThis milestone.
  2. 2021FLANSelf-Instruct was influenced by thisChapter 6. Teaching models what people want
  3. 2020GPT-3 Turning pointSelf-Instruct built directly on thisChapter 5. Scale becomes a science
  4. 2020Scaling laws Turning pointGPT-3 was influenced by thisChapter 5. Scale becomes a science
  5. 2019T5 Turning pointFLAN was influenced by thisChapter 4. Pretrain once, use everywhere
  6. 2019C4 Turning pointT5 was influenced by thisChapter 4. Pretrain once, use everywhere
  7. 2019Sparse Transformer Turning pointGPT-3 built directly on thisChapter 4. Pretrain once, use everywhere
  8. 2019GPT-2 Turning pointGPT-3 built directly on thisChapter 4. Pretrain once, use everywhere
  9. 2019WebTextGPT-2 was influenced by thisChapter 4. Pretrain once, use everywhere
  10. 2019Transformer-XLSparse Transformer was influenced by this (indirect influence)Chapter 4. Pretrain once, use everywhere
  11. 2018BERT Turning pointT5 was influenced by thisChapter 4. Pretrain once, use everywhere
  12. 2018SentencePieceT5 built directly on thisChapter 4. Pretrain once, use everywhere
  13. 2018GPT-1GPT-2 built directly on thisChapter 4. Pretrain once, use everywhere
  14. 2018ELMoBERT was influenced by this
    GPT-1 was influenced by this (indirect influence)    Chapter 4. Pretrain once, use everywhere
  15. 2018ULMFiTBERT was influenced by this
    GPT-1 was influenced by this (indirect influence)    Chapter 4. Pretrain once, use everywhere
  16. 2017CoVeELMo was influenced by thisChapter 3. Attention, then only attention
  17. 2017Transformer Turning pointGPT-2 built directly on this
    Sparse Transformer built directly on this
    T5 built directly on this    Chapter 3. Attention, then only attention
  18. 2017ConvS2STransformer was influenced by this (indirect influence)Chapter 3. Attention, then only attention
  19. 2016Layer normGPT-2 built directly on thisChapter 3. Attention, then only attention
  20. 2015Residual connectionsTransformer was influenced by thisChapter 3. Attention, then only attention
  21. 2015LM pretraining Turning pointBERT was influenced by this
    GPT-1 was influenced by this    Chapter 2. Words become vectors
  22. 2015BPE tokenizationGPT-2 built directly on thisChapter 2. Words become vectors
  23. 2015Luong attentionTransformer was influenced by thisChapter 3. Attention, then only attention
  24. 2014AdamTransformer was influenced by thisChapter 2. Words become vectors
  25. 2014GloVeELMo built directly on thisChapter 2. Words become vectors
  26. 2014Seq2seq Turning pointTransformer built directly on thisChapter 2. Words become vectors
  27. 2014Attention Turning pointTransformer built directly on thisChapter 3. Attention, then only attention
  28. 2014GRU / encoder-decoderAttention built directly on thisChapter 2. Words become vectors
  29. 2013word2vecELMo was influenced by thisChapter 2. Words become vectors
  30. 2011NLP from scratchELMo was influenced by thisChapter 1. Language as probability
  31. 2003Neural LMNLP from scratch was influenced by this
    word2vec was influenced by this    Chapter 1. Language as probability
  32. 1997LSTM Turning pointSeq2seq built directly on this
    ELMo built directly on this    Chapter 1. Language as probability
  33. 1990Recurrent networksLSTM was influenced by thisChapter 1. Language as probability

After: what it led to

Self-Instruct has 1 documented descendants across 1 chapter, reaching forward to Alpaca in 2023. Showing all of them.

Self-Instruct has 1 documented descendants across 1 chapter, reaching forward to Alpaca in 2023. Showing all of them.

Self-Instruct has 4 recorded descendants, including indirect ones, across 1 chapter, reaching forward to FineWeb in 2024. Showing its direct descendants.

Self-Instruct has 4 recorded descendants, including indirect ones, across 1 chapter, reaching forward to FineWeb in 2024. Showing all of them.

  1. 2022Self-InstructThis milestone.
  2. 2023Alpacabuilt directly on Self-InstructChapter 8. The weights get out
  1. 2022Self-InstructThis milestone.
  2. 2023Alpacabuilt directly on Self-InstructChapter 8. The weights get out
  3. 2023phi-1was influenced by Self-Instruct (indirect influence)Chapter 8. The weights get out
  4. 2024Phi-3built directly on phi-1Chapter 8. The weights get out
  5. 2024FineWebwas influenced by phi-1Chapter 8. The weights get out

Direct links and their sources

Where it came from

Built directly on

Influenced by

What it led to

Built on directly by

Influenced

Terms