Model · chatgpt
ChatGPT (Nov. 2022 launch model)
Developer: OpenAI
Availability: never_released · checked 2026-09-24 · source
Weights never released; API access status not tracked by this field.
Raw record: /data/models/chatgpt.json
Fields
- id
- chatgpt
- developer
- OpenAI
- release_date
- 2022-11-30recorded · source
- weights_status
- closed
- availability
- never_released · checked 2026-09-24 · sourcenote: Weights never released; API access status not tracked by this field.
- license
- Proprietary service; no weights released.partial · source
- architecture
- family
- decoder_only
- note
- Stub specimen. OpenAI: 'fine-tuned from a model in the GPT-3.5 series', trained with RLHF, 'a sibling model to InstructGPT'. The model behind the ChatGPT product changed over time; which version produced the ShareGPT / UltraChat data is not_recorded.
- n_layers
- nullnot_recorded
- hidden_size
- nullnot_recorded
- n_heads
- nullnot_recorded
- vocab_size
- nullnot_recorded
- positional_encoding
- nullnot_recorded
- training_data
- nullnot_recorded
- techniques
- rlhf
- primary_sources
- record_history
- date:2026-09-24 · change:created as a stub specimen to resolve vicuna/zephyr distilled_from_outputs edges · by:wilson-pruitt + claude ·date:2026-09-24 · change:availability checked and recorded · by:wilson-pruitt + claude ·
Parents
No edges recorded.
Children
Influence without weights
- ← distilled_from_outputs ShareGPT conversations (LMSYS collection) declared source ShareGPT.com is a site 'where users can share their ChatGPT conversations.' User-curated, not generated by LMSYS.
- ← distilled_from_outputs UltraChat declared source Builder's own paper: both sides of each dialogue generated by 'ChatGPT Turbo APIs'. Which ChatGPT snapshot is not_recorded.
- ← distilled_from_outputs OpenOrca declared source Card: '~3.2M GPT-3.5 completions'. Linked to the unversioned ChatGPT stub; the exact GPT-3.5 model is not_recorded on the card.
- ← distilled_from_outputs Baize self-chat data declared source Paper: 'leveraging ChatGPT (gpt-3.5-turbo)' for self-chat; snapshot not_recorded.
- ← distilled_from_outputs GPT4All prompt generations declared source Technical report: 'We collected roughly one million prompt-response pairs using the GPT-3.5-Turbo OpenAI API'; snapshot not_recorded.
- ← distilled_from_outputs WizardLM Evol-Instruct 70k declared source Builders' own paper: evolution 'using Azure OpenAI ChatGPT API. Then, we leverage ChatGPT to generate responses.' Which ChatGPT snapshot is not_recorded (paper says gpt-3.5-turbo from the Azure portal).
- ← distilled_from_outputs Code Evol-Instruct (WizardCoder) declared source Builders' own paper: 'OpenAI's gpt3.5turbo is used to evolve the dataset and generate responses.' Snapshot not_recorded.
- ← distilled_from_outputs GPT4All-J prompt generations declared source Builders' report: 'The assistant data was gathered from OpenAI's GPT3.5-Turbo'; the dataset 'is a superset of the original 400k points GPT4All dataset'. v1.3 also adds ShareGPT and Dolly rows, which this edge does not cover. Snapshot not_recorded.
- ← distilled_from_outputs Orca 1 explanation-tuning data (FLAN-5M / FLAN-1M) declared source Builder's own paper: 'We use Azure OpenAI API to collect ChatGPT (GPT-3.5-turbo) responses to FLAN-5M'. Snapshot not_recorded.
- ← feedback_from Baize SDF data (ChatGPT-ranked self-generations) declared source Paper: 'we use the resulted Baize v1.5 models to generate four responses for each instruction from the Quora dataset mentioned in Table 2. We then engage ChatGPT using the prompt provided in Appendix C to rank generate responses for self-distillation. Finally, we select the best response ranked by ChatGPT to finetune the model.' ChatGPT ranked candidates; the selected text is Baize v1.5's own output, so this is feedback (judgments), not distilled_from_outputs. ChatGPT snapshot not_recorded.
Read in
No station on the reading path has touched this record yet.