Model · pythia-chat-base-7b-v0-16
Pythia-Chat-Base-7B-v0.16
Developer: Together Computer (with LAION and Ontocord.ai)
Availability: available · checked 2026-09-24 · source
Raw record: /data/models/pythia-chat-base-7b-v0-16.json
Fields
- id
- pythia-chat-base-7b-v0-16
- identifiers
- huggingface
- togethercomputer/Pythia-Chat-Base-7B-v0.16
- developer
- Together Computer (with LAION and Ontocord.ai)
- release_date
- 2023-03-22partial · sourcenote: HF repo creation date; the card gives a version (v0.16) but no release date.
- weights_status
- open
- availability
- available · checked 2026-09-24 · source
- license
- apache-2.0recorded · sourcenote: Card: 'License: Apache 2.0'. Uploader is the developer.
- architecture
- family
- decoder_only
- note
- family read from config 'architectures': ['GPTNeoXForCausalLM'].
- n_layers
- 32recorded · source
- hidden_size
- 4096recorded · source
- n_heads
- 32recorded · source
- vocab_size
- 50432recorded · source
- context_length
- 2048recorded · source
- positional_encoding
- rotary (RoPE), partial: 25% of dimsrecorded · sourcenote: Propagated from pythia-6-9b-deduped, the base per the OpenChatKit README; config matches (32 layers, hidden 4096).
- training_data
- OIG-43M: 'a collection of 43 million high-quality instructions', curated by Together with LAION and Ontocord.ai (OpenChatKit models 'were trained on the OIG-43M training dataset'); plus 'a small amount of feedback data' (unspecified) in a further fine-tuning step.recorded · sourcenote: The feedback data is described only as 'a small amount'; its source is not_recorded.
- techniques
- primary_sources
- record_history
- date:2026-09-24 · change:ingested as candidate from HF (togethercomputer/Pythia-Chat-Base-7B-v0.16@97aa918c383820e1a69f042801091d7deb996c20) · by:ingest_hf.py ·date:2026-09-24 · change:preparer: developer, license, positional encoding (propagated from pythia-6-9b-deduped), training data, 2 edges prepared; sources: card, OpenChatKit README · by:claude (preparer, Sonnet 5) ·date:2026-09-24 · change:reviewed and promoted from staging (2 edge(s) accepted) · by:Wilson Pruitt ·
Parents
Weights descend
- fine_tuned_from → pythia-6.9b-deduped declared source OpenChatKit README: 'Pythia-Chat-Base-7B is a 7B-parameter fine-tuned variant of Pythia-6.9B-deduped from Eleuther AI'; the HF card says 'fine-tuned from EleutherAI's Pythia 7B' (no variant named). The README is what pins the deduped 6.9B.
Training data
- trained_on → Open Instruction Generalist (OIG) declared source README: 'The chat model was trained on the OIG dataset built by LAION, Together, and Ontocord.ai'; card: 43M-instruction collection (OIG-43M). The `oig` record covers the LAION OIG collection ('currently at 44M'); the 43M snapshot vs the current collection is not distinguished.
Children
No edges recorded.
Read in
No station on the reading path has touched this record yet.