Research Note

Llama 3.1 Technical Release Source Record

The July 23, 2024 release record should separate the paper, launch post, model card, pretrained weights, instruction-tuned weights, repository utilities, prompt format, l

Aug 4, 20261 min readBy Dalton Anderson
In this article

Llama 3.1 Technical Release Source Record

The July 23, 2024 release record should separate the paper, launch post, model card, pretrained weights, instruction-tuned weights, repository utilities, prompt format, license, acceptable-use policy, reported benchmarks, safeguards, and reference system.

The model card identifies 8B, 70B, and 405B text models, 128K context, supported languages, 15T-plus pretraining tokens, a December 2023 knowledge cutoff, intended uses, reported evaluation, and the custom Llama 3.1 Community License.

The license grants specified rights subject to attribution, redistribution, acceptable-use, and large-user commercial terms. Legal review, exact version, and intended use remain required.

Later Llama families are successors. They do not alter the identity, license, or evidence of Llama 3.1.

Sources

Follow the evidence.

  1. Introducing Llama 3.1ai.meta.com
  2. ai.meta.com: the llama 3 herd of modelsai.meta.com
  3. arxiv.org: 1810arxiv.org
  4. crfm.stanford.edu: indexcrfm.stanford.edu
  5. arxiv.org: 2203arxiv.org
  6. open.spotify.com: 0iRBPcPw9iYjpUVAVWSkRCopen.spotify.com
  7. NIST AI Risk Management Frameworknist.gov
  8. github.com: MODEL CARDgithub.com
  9. daltonanderson.ghost.io: metas llama 3 1 inside the ai research paperdaltonanderson.ghost.io
  10. Meta Llama models repositorygithub.com
  11. arxiv.org: 2001arxiv.org
  12. youtu.be: UMhmWCor1kYyoutu.be
  13. github.com: LICENSEgithub.com

From this episode

Two useful next steps.

Research Note · 1 min

Training Data Disclosure Evidence Framework

Classify each statement as exact item disclosure, named dataset, source category, proportion, collection method, filtering method, deduplication method, token count, time

Research Note · 1 min

Scaling Law Decision Framework

Scaling laws are empirical relationships estimated from a defined model family, dataset regime, metric, compute range, and training procedure. They can guide allocation a

Return to the episode
Llama 3.1 Technical Release Source Record