{
  "slug": "llama-3-8b",
  "order": 1,
  "name": "Llama 3 (8B)",
  "baseTitle": "Llama 3 8B",
  "company": "Meta",
  "country": "United States",
  "flag": "🇺🇸",
  "date": "2024-04-18",
  "decoderType": "Dense",
  "attention": "GQA with RoPE",
  "layerMix": "32 GQA",
  "scale": "8B parameters",
  "totalParams": 8000000000,
  "activeParams": 8000000000,
  "context": "8,192",
  "contextTokens": 8192,
  "license": "Llama 3 Community License Agreement",
  "kvCacheText": "128 KiB",
  "kvCacheKiB": 128,
  "kvCacheLevel": "Moderate",
  "summary": "Reference dense Llama stack used to contrast OLMo 2's normalization and attention choices.",
  "keyDetail": "Pre-norm baseline; wider than OLMo 2 at a similar scale.",
  "image": {
    "full": "https://sebastianraschka.com/llm-architecture-gallery/images/architectures/llama-3-8b.webp",
    "thumb": "https://sebastianraschka.com/llm-architecture-gallery/images/architectures/thumbnails/llama-3-8b.webp",
    "alt": "Figure 10: Llama 3 8B",
    "width": 914,
    "height": 960
  },
  "articleUrl": "https://magazine.sebastianraschka.com/p/the-big-llm-architecture-comparison#%C2%A723-olmo-2-summary",
  "links": [
    {
      "label": "View in article",
      "href": "https://magazine.sebastianraschka.com/p/the-big-llm-architecture-comparison#%C2%A723-olmo-2-summary",
      "title": null
    },
    {
      "label": "From scratch",
      "href": "https://github.com/rasbt/LLMs-from-scratch/tree/main/ch05/07_gpt_to_llama",
      "title": null
    },
    {
      "label": "config.json",
      "href": "https://huggingface.co/meta-llama/Meta-Llama-3-8B/blob/main/config.json",
      "title": "meta-llama/Meta-Llama-3-8B"
    },
    {
      "label": "License",
      "href": "https://huggingface.co/meta-llama/Meta-Llama-3-8B/blob/main/LICENSE",
      "title": null
    },
    {
      "label": "Tech report",
      "href": "https://arxiv.org/pdf/2407.21783",
      "title": null
    }
  ],
  "concepts": [
    {
      "slug": "gqa",
      "label": "GQA",
      "name": "Grouped-Query Attention",
      "url": "https://llm-atlas-7or.pages.dev/concepts/gqa/"
    }
  ],
  "aai": {
    "v4": {
      "label": "AA Intelligence Index v4.3.2",
      "total": 4.8,
      "profile": {
        "Scientific": 2.5
      },
      "url": "https://artificialanalysis.ai/models/llama-3-instruct-8b"
    }
  },
  "sourceUrl": "https://sebastianraschka.com/llm-architecture-gallery/#card-llama-3-8b",
  "family": "dense",
  "url": "https://llm-atlas-7or.pages.dev/models/llama-3-8b/",
  "markdownUrl": "https://llm-atlas-7or.pages.dev/models/llama-3-8b.md"
}