Dataset · gpt4all

GPT4All prompt generations

Builder: Nomic AI
Availability: available · checked 2026-09-24 · source

Raw record: /data/datasets/gpt4all.json

Fields

id

gpt4all

identifiers

huggingface

nomic-ai/gpt4all_prompt_generations

builder

Nomic AI

release_date

2023-03-27partial · source

note:  HF dataset repo creation date, not the builder's own release announcement.

availability

available · checked 2026-09-24 · source

content

Technical report: 'roughly one million prompt-response pairs using the GPT-3.5-Turbo OpenAI API between March 20, 2023 and March 26th, 2023', prompts drawn from LAION OIG (unified chip2), Stack Overflow questions, and Bigscience/P3; P3 was later dropped, leaving a final subset of 437,605 prompt-generation pairs used to train the first release.recorded · source

note:  Conflicting third-party statement: the StableLM-Tuned-Alpha card describes GPT4All Prompt Generations as '400k prompts and responses generated by GPT-4' (https://huggingface.co/stabilityai/stablelm-tuned-alpha-7b). The builder's own report says GPT-3.5-Turbo throughout; the builder's wording is kept, and the edge below rests on it.

primary_sources

https://huggingface.co/datasets/nomic-ai/gpt4all_prompt_generations

https://s3.amazonaws.com/static.nomic.ai/gpt4all/2023_GPT4All_Technical_Report.pdf

record_history

date:2026-09-24 · change:created from primary sources (tranche 2 dataset records) · by:wilson-pruitt + claude ·

date:2026-09-24 · change:note added reconciling the GPT-3.5-Turbo (builder) vs GPT-4 (StableLM card) generator claims · by:wilson-pruitt + claude ·

Parents

Influence without weights

• distilled_from_outputs → ChatGPT (Nov. 2022 launch model) declared source Technical report: 'We collected roughly one million prompt-response pairs using the GPT-3.5-Turbo OpenAI API'; snapshot not_recorded.

Children

Training data

• ← trained_on GPT4All-J declared source Report: 'we curated the GPT4All-J dataset by augmenting the original 400k GPT4All examples' and the 800k set is 'a superset of the original 400k points GPT4All dataset'. This is the distillation path: the GPT4All examples were generated with GPT-3.5-Turbo (see the `gpt4all` dataset record).
• ← trained_on StableLM-Tuned-Alpha-7B declared source Card lists 'GPT4All Prompt Generations' and describes it as 'generated by GPT-4'. The GPT4All technical report (see the `gpt4all` record) says GPT-3.5-Turbo; the edge follows the dataset name, and the card's model attribution is flagged, not adopted.

Read in

No station on the reading path has touched this record yet.