{
  "id": "cerebras-gpt-13b",
  "name": "Cerebras-GPT 13B",
  "identifiers": {
    "huggingface": [
      "cerebras/Cerebras-GPT-13B"
    ]
  },
  "developer": "Cerebras Systems (Nolan Dey, Gurpreet Gosal, Zhiming (Charles) Chen, Hemant Khachane, William Marshall, Ribhu Pathria, Marvin Tom, Joel Hestness; trained on the Cerebras Wafer-Scale Cluster)",
  "release_date": {
    "value": "2023-03-28",
    "source": "https://www.cerebras.net/blog/cerebras-gpt-a-family-of-open-compute-efficient-large-language-models/",
    "status": "recorded",
    "note": "Cerebras blog post dated Mar 28 2023; paper arXiv 2304.03208 v1 is dated 6 Apr 2023."
  },
  "weights_status": "unknown",
  "availability": {
    "value": "removed",
    "checked": "2026-09-24",
    "source": "https://huggingface.co/cerebras/Cerebras-GPT-13B",
    "note": "HF API returned HTTP 401."
  },
  "license": {
    "value": "Apache-2.0",
    "source": "https://www.cerebras.net/blog/cerebras-gpt-a-family-of-open-compute-efficient-large-language-models/",
    "status": "recorded",
    "note": "Blog: 'All models, weights, and checkpoints are available on Hugging Face and GitHub under the Apache 2.0 license.' HF repo unreachable (HTTP 401) so the card was not read."
  },
  "architecture": {
    "family": "decoder_only",
    "note": "HF config not retrieved (repo unreachable: HTTP 401). Values from paper Table 1 (13B row: d_model 5120, n_layers 40, d_head 128, d_ffn 20480) and section 2.1. n_heads is derived (5120 / 128 = 40); the paper gives d_head, not the head count. Positional encoding is not stated in the sections read.",
    "n_layers": {
      "value": 40,
      "source": "https://arxiv.org/abs/2304.03208",
      "status": "recorded",
      "note": "Paper Table 1."
    },
    "hidden_size": {
      "value": 5120,
      "source": "https://arxiv.org/abs/2304.03208",
      "status": "recorded",
      "note": "Paper Table 1 d_model."
    },
    "n_heads": {
      "value": 40,
      "source": "https://arxiv.org/abs/2304.03208",
      "status": "recorded",
      "note": "Derived: d_model 5120 / d_head 128; paper Table 1 does not list the count directly."
    },
    "vocab_size": {
      "value": 50257,
      "source": "https://arxiv.org/abs/2304.03208",
      "status": "recorded",
      "note": "Paper sec. 2.2: GPT-2 BPE vocabulary of size 50257."
    },
    "context_length": {
      "value": 2048,
      "source": "https://arxiv.org/abs/2304.03208",
      "status": "recorded",
      "note": "Paper sec. 2.1: 'maximum sequence length of 2048 tokens'."
    },
    "positional_encoding": {
      "value": null,
      "status": "not_recorded"
    }
  },
  "training_data": {
    "value": "The Pile (Gao et al., 2020), no deduplication; 257.1B tokens (about 20 tokens per parameter, Chinchilla-optimal).",
    "source": "https://arxiv.org/abs/2304.03208",
    "status": "recorded",
    "note": "Paper: 'We train Cerebras-GPT models on the Eleuther Pile dataset following DeepMind Chinchilla scaling rules'; 'We do not perform deduplication of Pile'; Table 1 total tokens 257.1B for 13B."
  },
  "techniques": [],
  "primary_sources": [
    "https://huggingface.co/cerebras/Cerebras-GPT-13B"
  ],
  "record_history": [
    {
      "date": "2026-09-24",
      "change": "ingested as candidate from HF (cerebras/Cerebras-GPT-13B@None)",
      "by": "ingest_hf.py"
    },
    {
      "date": "2026-09-24",
      "change": "preparer: developer, release date (blog), license (blog), architecture (paper), training data, 2 edges (trained_on the-pile; design_follows gpt-3); sources: arXiv 2304.03208, Cerebras blog",
      "by": "claude (preparer, Sonnet 5)"
    },
    {
      "date": "2026-09-24",
      "change": "reviewed and promoted from staging (2 edge(s) accepted)",
      "by": "Wilson Pruitt"
    }
  ]
}