{
  "reviewed": "2026-11-13",
  "sources": [
    {
      "title": "Neural Machine Translation of Rare Words with Subword Units",
      "url": "https://aclanthology.org/P16-1162/",
      "author": "Rico Sennrich, Barry Haddow, Alexandra Birch",
      "published": "2016-08",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, code, figures, or video assets reproduced.",
      "rights": "Linked reference; copyright remains with its authors. Original explanations and examples created for this site.",
      "lessons": [
        "tokens",
        "tokenization"
      ]
    },
    {
      "title": "Embedding — PyTorch 2.15 documentation",
      "url": "https://docs.pytorch.org/docs/2.15/generated/torch.nn.Embedding.html",
      "author": "PyTorch contributors",
      "published": "Version 2.15; accessed 2026-11-13",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, code, figures, or video assets reproduced.",
      "rights": "Linked reference; copyright remains with its authors. Original explanations and examples created for this site.",
      "lessons": [
        "tokens",
        "embeddings"
      ]
    },
    {
      "title": "Neural Networks: Zero to Hero",
      "url": "https://github.com/karpathy/nn-zero-to-hero",
      "author": "Andrej Karpathy",
      "published": "Living course; accessed 2026-11-13",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, code, figures, or video assets reproduced.",
      "rights": "Linked reference; copyright remains with its authors. Original explanations and examples created for this site.",
      "lessons": [
        "tokenization",
        "neural-networks"
      ]
    },
    {
      "title": "SentencePiece",
      "url": "https://aclanthology.org/D18-2012/",
      "author": "Taku Kudo, John Richardson",
      "published": "2018-11",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, code, figures, or video assets reproduced.",
      "rights": "Linked reference; copyright remains with its authors. Original explanations and examples created for this site.",
      "lessons": [
        "tokenization"
      ]
    },
    {
      "title": "minbpe: byte-pair encoding reference implementation",
      "url": "https://github.com/karpathy/minbpe",
      "author": "Andrej Karpathy",
      "published": "Living repository; accessed 2026-11-13",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, code, figures, or video assets reproduced.",
      "rights": "Linked reference; copyright remains with its authors. Original explanations and examples created for this site.",
      "lessons": [
        "tokenization"
      ]
    },
    {
      "title": "The Pile: An 800GB Dataset of Diverse Text for Language Modeling",
      "url": "https://arxiv.org/abs/2101.00027",
      "author": "Leo Gao et al., EleutherAI",
      "published": "2020-12-31",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, code, figures, or video assets reproduced.",
      "rights": "Linked reference; copyright remains with its authors. Original explanations and examples created for this site.",
      "lessons": [
        "training-data"
      ]
    },
    {
      "title": "Deduplicating Training Data Makes Language Models Better",
      "url": "https://arxiv.org/abs/2107.06499",
      "author": "Katherine Lee et al.",
      "published": "2022-03-24 (v2)",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, code, figures, or video assets reproduced.",
      "rights": "Linked reference; copyright remains with its authors. Original explanations and examples created for this site.",
      "lessons": [
        "data-quality"
      ]
    },
    {
      "title": "Efficient Estimation of Word Representations in Vector Space",
      "url": "https://arxiv.org/abs/1301.3781",
      "author": "Tomas Mikolov et al.",
      "published": "2013-09-07 (v3)",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, code, figures, or video assets reproduced.",
      "rights": "Linked reference; copyright remains with its authors. Original explanations and examples created for this site.",
      "lessons": [
        "embeddings"
      ]
    },
    {
      "title": "Attention Is All You Need",
      "url": "https://arxiv.org/abs/1706.03762",
      "author": "Ashish Vaswani et al., Google and University of Toronto",
      "published": "2017; revised 2023-08-02 (v7)",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, code, figures, or video assets reproduced.",
      "rights": "Linked reference; copyright remains with its authors. Original explanations and examples created for this site.",
      "lessons": [
        "positions",
        "transformers",
        "attention-heads"
      ]
    },
    {
      "title": "RoFormer: Enhanced Transformer with Rotary Position Embedding",
      "url": "https://arxiv.org/abs/2104.09864",
      "author": "Jianlin Su et al.",
      "published": "2021; revised 2023-11-08 (v5)",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, code, figures, or video assets reproduced.",
      "rights": "Linked reference; copyright remains with its authors. Original explanations and examples created for this site.",
      "lessons": [
        "positions"
      ]
    },
    {
      "title": "Linear — PyTorch 2.15 documentation",
      "url": "https://docs.pytorch.org/docs/2.15/generated/torch.nn.Linear.html",
      "author": "PyTorch contributors",
      "published": "Version 2.15; accessed 2026-11-13",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, code, figures, or video assets reproduced.",
      "rights": "Linked reference; copyright remains with its authors. Original explanations and examples created for this site.",
      "lessons": [
        "parameters",
        "neural-networks"
      ]
    },
    {
      "title": "nanoGPT: historical GPT implementation",
      "url": "https://github.com/karpathy/nanoGPT",
      "author": "Andrej Karpathy and contributors",
      "published": "Historical implementation; repository accessed 2026-11-13",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, code, figures, or video assets reproduced.",
      "rights": "Linked reference; copyright remains with its authors. Original explanations and examples created for this site.",
      "lessons": [
        "parameters",
        "transformers"
      ]
    },
    {
      "title": "Layer Normalization",
      "url": "https://arxiv.org/abs/1607.06450",
      "author": "Jimmy Lei Ba, Jamie Ryan Kiros, Geoffrey E. Hinton",
      "published": "2016-07-21",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, code, figures, or video assets reproduced.",
      "rights": "Linked reference; copyright remains with its authors. Original explanations and examples created for this site.",
      "lessons": [
        "transformers",
        "residual-stream"
      ]
    },
    {
      "title": "Scaled dot-product attention — PyTorch 2.15 documentation",
      "url": "https://docs.pytorch.org/docs/2.15/generated/torch.nn.functional.scaled_dot_product_attention.html",
      "author": "PyTorch contributors",
      "published": "Version 2.15; accessed 2026-11-13",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, code, figures, or video assets reproduced.",
      "rights": "Linked reference; copyright remains with its authors. Original explanations and examples created for this site.",
      "lessons": [
        "attention",
        "causal-masking"
      ]
    },
    {
      "title": "Attention is not Explanation",
      "url": "https://arxiv.org/abs/1902.10186",
      "author": "Sarthak Jain, Byron C. Wallace",
      "published": "2019-05-08 (v3)",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, code, figures, or video assets reproduced.",
      "rights": "Linked reference; copyright remains with its authors. Original explanations and examples created for this site.",
      "lessons": [
        "attention"
      ]
    },
    {
      "title": "Are Sixteen Heads Really Better than One?",
      "url": "https://arxiv.org/abs/1905.10650",
      "author": "Paul Michel, Omer Levy, Graham Neubig",
      "published": "2019-11-04 (v3)",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, code, figures, or video assets reproduced.",
      "rights": "Linked reference; copyright remains with its authors. Original explanations and examples created for this site.",
      "lessons": [
        "attention-heads"
      ]
    },
    {
      "title": "GQA: Training Generalized Multi-Query Transformer Models from Multi-Head Checkpoints",
      "url": "https://arxiv.org/abs/2305.13245",
      "author": "Joshua Ainslie et al.",
      "published": "2023-12-23 (v3)",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, code, figures, or video assets reproduced.",
      "rights": "Linked reference; copyright remains with its authors. Original explanations and examples created for this site.",
      "lessons": [
        "attention-heads",
        "kv-cache"
      ]
    },
    {
      "title": "GLU Variants Improve Transformer",
      "url": "https://arxiv.org/abs/2002.05202",
      "author": "Noam Shazeer",
      "published": "2020-02-12",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, code, figures, or video assets reproduced.",
      "rights": "Linked reference; copyright remains with its authors. Original explanations and examples created for this site.",
      "lessons": [
        "feed-forward"
      ]
    },
    {
      "title": "Transformer Feed-Forward Layers Are Key-Value Memories",
      "url": "https://arxiv.org/abs/2012.14913",
      "author": "Mor Geva, Roei Schuster, Jonathan Berant, Omer Levy",
      "published": "2021-09-05 (v2)",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, code, figures, or video assets reproduced.",
      "rights": "Linked reference; copyright remains with its authors. Original explanations and examples created for this site.",
      "lessons": [
        "feed-forward"
      ]
    },
    {
      "title": "Deep Residual Learning for Image Recognition",
      "url": "https://arxiv.org/abs/1512.03385",
      "author": "Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun",
      "published": "2015-12-10",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, code, figures, or video assets reproduced.",
      "rights": "Linked reference; copyright remains with its authors. Original explanations and examples created for this site.",
      "lessons": [
        "residual-stream"
      ]
    },
    {
      "title": "A Mathematical Framework for Transformer Circuits",
      "url": "https://transformer-circuits.pub/2021/framework/index.html",
      "author": "Nelson Elhage et al., Anthropic",
      "published": "2021-12-22",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, code, figures, or video assets reproduced.",
      "rights": "Linked reference; copyright remains with its authors. Original explanations and examples created for this site.",
      "lessons": [
        "residual-stream"
      ]
    },
    {
      "title": "Using the Output Embedding to Improve Language Models",
      "url": "https://arxiv.org/abs/1608.05859",
      "author": "Ofir Press, Lior Wolf",
      "published": "2017-02-21 (v3)",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, code, figures, or video assets reproduced.",
      "rights": "Linked reference; copyright remains with its authors. Original explanations and examples created for this site.",
      "lessons": [
        "logits"
      ]
    },
    {
      "title": "Softmax — PyTorch 2.15 documentation",
      "url": "https://docs.pytorch.org/docs/2.15/generated/torch.nn.Softmax.html",
      "author": "PyTorch contributors",
      "published": "Version 2.15; accessed 2026-11-13",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, code, figures, or video assets reproduced.",
      "rights": "Linked reference; copyright remains with its authors. Original explanations and examples created for this site.",
      "lessons": [
        "logits",
        "softmax"
      ]
    },
    {
      "title": "On Calibration of Modern Neural Networks",
      "url": "https://arxiv.org/abs/1706.04599",
      "author": "Chuan Guo, Geoff Pleiss, Yu Sun, Kilian Q. Weinberger",
      "published": "2017-08-03 (v2)",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, code, figures, or video assets reproduced.",
      "rights": "Linked reference; copyright remains with its authors. Original explanations and examples created for this site.",
      "lessons": [
        "softmax"
      ]
    },
    {
      "title": "Generation configuration — Transformers documentation",
      "url": "https://huggingface.co/docs/transformers/main/en/main_classes/text_generation",
      "author": "Hugging Face contributors",
      "published": "Main documentation; accessed 2026-11-13",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, code, figures, or video assets reproduced.",
      "rights": "Linked reference; copyright remains with its authors. Original explanations and examples created for this site.",
      "lessons": [
        "temperature"
      ]
    },
    {
      "title": "Generation strategies — Transformers documentation",
      "url": "https://huggingface.co/docs/transformers/main/en/generation_strategies",
      "author": "Hugging Face contributors",
      "published": "Main documentation; accessed 2026-11-13",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, code, figures, or video assets reproduced.",
      "rights": "Linked reference; copyright remains with its authors. Original explanations and examples created for this site.",
      "lessons": [
        "sampling",
        "autoregression"
      ]
    },
    {
      "title": "The Curious Case of Neural Text Degeneration",
      "url": "https://arxiv.org/abs/1904.09751",
      "author": "Ari Holtzman et al.",
      "published": "2020-02-14 (v2)",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, code, figures, or video assets reproduced.",
      "rights": "Linked reference; copyright remains with its authors. Original explanations and examples created for this site.",
      "lessons": [
        "sampling"
      ]
    },
    {
      "title": "Language Models are Few-Shot Learners",
      "url": "https://arxiv.org/abs/2005.14165",
      "author": "Tom B. Brown et al., OpenAI",
      "published": "2020-07-22 (v4)",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, code, figures, or video assets reproduced.",
      "rights": "Linked reference; copyright remains with its authors. Original explanations and examples created for this site.",
      "lessons": [
        "next-token-prediction",
        "pretraining",
        "in-context-learning"
      ]
    },
    {
      "title": "How caching works — Transformers documentation",
      "url": "https://huggingface.co/docs/transformers/main/en/cache_explanation",
      "author": "Hugging Face contributors",
      "published": "Main documentation; accessed 2026-11-13",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, code, figures, or video assets reproduced.",
      "rights": "Linked reference; copyright remains with its authors. Original explanations and examples created for this site.",
      "lessons": [
        "autoregression",
        "kv-cache"
      ]
    },
    {
      "title": "Lost in the Middle: How Language Models Use Long Contexts",
      "url": "https://arxiv.org/abs/2307.03172",
      "author": "Nelson F. Liu et al.",
      "published": "2023-11-20 (v3)",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, code, figures, or video assets reproduced.",
      "rights": "Linked reference; copyright remains with its authors. Original explanations and examples created for this site.",
      "lessons": [
        "context-windows",
        "context-management"
      ]
    },
    {
      "title": "Fast Transformer Decoding: One Write-Head is All You Need",
      "url": "https://arxiv.org/abs/1911.02150",
      "author": "Noam Shazeer",
      "published": "2019-11-06",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, code, figures, or video assets reproduced.",
      "rights": "Linked reference; copyright remains with its authors. Original explanations and examples created for this site.",
      "lessons": [
        "kv-cache"
      ]
    },
    {
      "title": "CrossEntropyLoss documentation",
      "url": "https://docs.pytorch.org/docs/stable/generated/torch.nn.CrossEntropyLoss.html",
      "author": "PyTorch contributors",
      "published": "Living documentation; checked 2026-11-13",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, graphics, code, or assets reproduced. Original explanation and examples.",
      "rights": "Linked reference. Copyright remains with its authors; no reproduction license relied upon.",
      "lessons": [
        "loss"
      ]
    },
    {
      "title": "Deep Learning, chapter 8: Optimization",
      "url": "https://www.deeplearningbook.org/contents/optimization.html",
      "author": "Ian Goodfellow, Yoshua Bengio and Aaron Courville",
      "published": "2016",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, graphics, code, or assets reproduced. Original explanation and examples.",
      "rights": "Linked reference. Copyright remains with its authors; no reproduction license relied upon.",
      "lessons": [
        "loss",
        "gradient-descent",
        "generalization"
      ]
    },
    {
      "title": "Automatic differentiation with torch.autograd",
      "url": "https://docs.pytorch.org/tutorials/beginner/basics/autogradqs_tutorial.html",
      "author": "PyTorch contributors",
      "published": "Living documentation; checked 2026-11-13",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, graphics, code, or assets reproduced. Original explanation and examples.",
      "rights": "Linked reference. Copyright remains with its authors; no reproduction license relied upon.",
      "lessons": [
        "backpropagation"
      ]
    },
    {
      "title": "Scaling Laws for Neural Language Models",
      "url": "https://arxiv.org/abs/2001.08361",
      "author": "Jared Kaplan et al.",
      "published": "2020-01-23",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, graphics, code, or assets reproduced. Original explanation and examples.",
      "rights": "Linked reference. Copyright remains with its authors; no reproduction license relied upon.",
      "lessons": [
        "scaling-laws"
      ]
    },
    {
      "title": "Training Compute-Optimal Large Language Models",
      "url": "https://arxiv.org/abs/2203.15556",
      "author": "Jordan Hoffmann et al.",
      "published": "2022-03-29",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, graphics, code, or assets reproduced. Original explanation and examples.",
      "rights": "Linked reference. Copyright remains with its authors; no reproduction license relied upon.",
      "lessons": [
        "scaling-laws",
        "compute"
      ]
    },
    {
      "title": "Predictability and Surprise in Large Generative Models",
      "url": "https://arxiv.org/abs/2202.07785",
      "author": "Deep Ganguli et al.",
      "published": "2022-02-15",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, graphics, code, or assets reproduced. Original explanation and examples.",
      "rights": "Linked reference. Copyright remains with its authors; no reproduction license relied upon.",
      "lessons": [
        "scaling-laws",
        "generalization"
      ]
    },
    {
      "title": "Training language models to follow instructions with human feedback",
      "url": "https://arxiv.org/abs/2203.02155",
      "author": "Long Ouyang et al.",
      "published": "2022-03-04",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, graphics, code, or assets reproduced. Original explanation and examples.",
      "rights": "Linked reference. Copyright remains with its authors; no reproduction license relied upon.",
      "lessons": [
        "supervised-finetuning",
        "rlhf"
      ]
    },
    {
      "title": "Multitask Prompted Training Enables Zero-Shot Task Generalization",
      "url": "https://arxiv.org/abs/2110.08207",
      "author": "Victor Sanh et al.",
      "published": "2021-10-15",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, graphics, code, or assets reproduced. Original explanation and examples.",
      "rights": "Linked reference. Copyright remains with its authors; no reproduction license relied upon.",
      "lessons": [
        "supervised-finetuning",
        "generalization"
      ]
    },
    {
      "title": "Deep reinforcement learning from human preferences",
      "url": "https://arxiv.org/abs/1706.03741",
      "author": "Paul Christiano et al.",
      "published": "2017-06-12",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, graphics, code, or assets reproduced. Original explanation and examples.",
      "rights": "Linked reference. Copyright remains with its authors; no reproduction license relied upon.",
      "lessons": [
        "preference-learning"
      ]
    },
    {
      "title": "Constitutional AI: Harmlessness from AI Feedback",
      "url": "https://arxiv.org/abs/2212.08073",
      "author": "Yuntao Bai et al.",
      "published": "2022-12-15",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, graphics, code, or assets reproduced. Original explanation and examples.",
      "rights": "Linked reference. Copyright remains with its authors; no reproduction license relied upon.",
      "lessons": [
        "preference-learning"
      ]
    },
    {
      "title": "Proximal Policy Optimization Algorithms",
      "url": "https://arxiv.org/abs/1707.06347",
      "author": "John Schulman et al.",
      "published": "2017-07-20",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, graphics, code, or assets reproduced. Original explanation and examples.",
      "rights": "Linked reference. Copyright remains with its authors; no reproduction license relied upon.",
      "lessons": [
        "rlhf"
      ]
    },
    {
      "title": "Direct Preference Optimization: Your Language Model is Secretly a Reward Model",
      "url": "https://arxiv.org/abs/2305.18290",
      "author": "Rafael Rafailov et al.",
      "published": "2023-05-29; revision 2024-07-29",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, graphics, code, or assets reproduced. Original explanation and examples.",
      "rights": "Linked reference. Copyright remains with its authors; no reproduction license relied upon.",
      "lessons": [
        "dpo"
      ]
    },
    {
      "title": "Llama 2: Open Foundation and Fine-Tuned Chat Models",
      "url": "https://arxiv.org/abs/2307.09288",
      "author": "Hugo Touvron et al.",
      "published": "2023-07-18",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, graphics, code, or assets reproduced. Original explanation and examples.",
      "rights": "Linked reference. Copyright remains with its authors; no reproduction license relied upon.",
      "lessons": [
        "base-vs-assistant"
      ]
    },
    {
      "title": "Chat templates",
      "url": "https://huggingface.co/docs/transformers/chat_templating",
      "author": "Hugging Face contributors",
      "published": "Living documentation; checked 2026-11-13",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, graphics, code, or assets reproduced. Original explanation and examples.",
      "rights": "Linked reference. Copyright remains with its authors; no reproduction license relied upon.",
      "lessons": [
        "prompt-structure"
      ]
    },
    {
      "title": "Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks",
      "url": "https://arxiv.org/abs/2005.11401",
      "author": "Patrick Lewis et al.",
      "published": "2020-05-22",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, graphics, code, or assets reproduced. Original explanation and examples.",
      "rights": "Linked reference. Copyright remains with its authors; no reproduction license relied upon.",
      "lessons": [
        "retrieval"
      ]
    },
    {
      "title": "Toolformer: Language Models Can Teach Themselves to Use Tools",
      "url": "https://arxiv.org/abs/2302.04761",
      "author": "Timo Schick et al.",
      "published": "2023-02-09",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, graphics, code, or assets reproduced. Original explanation and examples.",
      "rights": "Linked reference. Copyright remains with its authors; no reproduction license relied upon.",
      "lessons": [
        "tool-use"
      ]
    },
    {
      "title": "WebGPT: Browser-assisted question-answering with human feedback",
      "url": "https://arxiv.org/abs/2112.09332",
      "author": "Reiichiro Nakano et al.",
      "published": "2021-12-17",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, graphics, code, or assets reproduced. Original explanation and examples.",
      "rights": "Linked reference. Copyright remains with its authors; no reproduction license relied upon.",
      "lessons": [
        "tool-use"
      ]
    },
    {
      "title": "TruthfulQA: Measuring How Models Mimic Human Falsehoods",
      "url": "https://arxiv.org/abs/2109.07958",
      "author": "Stephanie Lin, Jacob Hilton and Owain Evans",
      "published": "2021-09-08",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, graphics, code, or assets reproduced. Original explanation and examples.",
      "rights": "Linked reference. Copyright remains with its authors; no reproduction license relied upon.",
      "lessons": [
        "hallucinations"
      ]
    },
    {
      "title": "Holistic Evaluation of Language Models",
      "url": "https://arxiv.org/abs/2211.09110",
      "author": "Percy Liang et al.",
      "published": "2022-11-16",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, graphics, code, or assets reproduced. Original explanation and examples.",
      "rights": "Linked reference. Copyright remains with its authors; no reproduction license relied upon.",
      "lessons": [
        "evaluation"
      ]
    },
    {
      "title": "Red Teaming Language Models to Reduce Harms",
      "url": "https://arxiv.org/abs/2209.07858",
      "author": "Deep Ganguli et al.",
      "published": "2022-08-23",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, graphics, code, or assets reproduced. Original explanation and examples.",
      "rights": "Linked reference. Copyright remains with its authors; no reproduction license relied upon.",
      "lessons": [
        "evaluation"
      ]
    },
    {
      "title": "Extracting Training Data from Large Language Models",
      "url": "https://arxiv.org/abs/2012.07805",
      "author": "Nicholas Carlini et al.",
      "published": "2020-12-14",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, graphics, code, or assets reproduced. Original explanation and examples.",
      "rights": "Linked reference. Copyright remains with its authors; no reproduction license relied upon.",
      "lessons": [
        "memorization"
      ]
    },
    {
      "title": "Deduplicating Training Data Mitigates Privacy Risks in Language Models",
      "url": "https://arxiv.org/abs/2202.06539",
      "author": "Nikhil Kandpal, Eric Wallace and Colin Raffel",
      "published": "2022-02-14",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, graphics, code, or assets reproduced. Original explanation and examples.",
      "rights": "Linked reference. Copyright remains with its authors; no reproduction license relied upon.",
      "lessons": [
        "memorization"
      ]
    },
    {
      "title": "FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awareness",
      "url": "https://arxiv.org/abs/2205.14135",
      "author": "Tri Dao et al.",
      "published": "2022-05-27",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, graphics, code, or assets reproduced. Original explanation and examples.",
      "rights": "Linked reference. Copyright remains with its authors; no reproduction license relied upon.",
      "lessons": [
        "compute"
      ]
    },
    {
      "title": "DeepSpeed Inference",
      "url": "https://arxiv.org/abs/2207.00032",
      "author": "Reza Yazdani Aminabadi et al.",
      "published": "2022-06-30",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, graphics, code, or assets reproduced. Original explanation and examples.",
      "rights": "Linked reference. Copyright remains with its authors; no reproduction license relied upon.",
      "lessons": [
        "compute"
      ]
    },
    {
      "title": "Efficient Memory Management for Large Language Model Serving with PagedAttention",
      "url": "https://arxiv.org/abs/2309.06180",
      "author": "Woosuk Kwon et al.",
      "published": "2023-09-12",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, graphics, code, or assets reproduced. Original explanation and examples.",
      "rights": "Linked reference. Copyright remains with its authors; no reproduction license relied upon.",
      "lessons": [
        "latency"
      ]
    },
    {
      "title": "GPTQ: Accurate Post-Training Quantization for Generative Pre-trained Transformers",
      "url": "https://arxiv.org/abs/2210.17323",
      "author": "Elias Frantar et al.",
      "published": "2022-10-31",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, graphics, code, or assets reproduced. Original explanation and examples.",
      "rights": "Linked reference. Copyright remains with its authors; no reproduction license relied upon.",
      "lessons": [
        "quantization"
      ]
    },
    {
      "title": "AWQ: Activation-aware Weight Quantization for LLM Compression and Acceleration",
      "url": "https://arxiv.org/abs/2306.00978",
      "author": "Ji Lin et al.",
      "published": "2023-06-01; accessed revision 2026-04-25",
      "accessed": "2026-11-13",
      "usage": "Reference only; no source text, graphics, code, or assets reproduced. Original explanation and examples.",
      "rights": "Linked reference. Copyright remains with its authors; no reproduction license relied upon.",
      "lessons": [
        "quantization"
      ]
    },
    {
      "title": "Deep Dive into LLMs like ChatGPT",
      "url": "https://www.youtube.com/watch?v=7xTGNNLPyMI",
      "author": "Andrej Karpathy",
      "published": "2025",
      "usage": "Pedagogical reference; external link only. No transcript, footage, voice, likeness, or graphics reproduced.",
      "rights": "YouTube-hosted copyrighted video; no reuse permission assumed.",
      "accessed": "2026-11-13",
      "lessons": []
    },
    {
      "title": "[1hr Talk] Intro to Large Language Models",
      "url": "https://www.youtube.com/watch?v=zjkBMFhNj_g",
      "author": "Andrej Karpathy",
      "published": "2023",
      "usage": "Shorter companion reference; external link only.",
      "rights": "YouTube-hosted copyrighted video; no reuse permission assumed.",
      "accessed": "2026-11-13",
      "lessons": []
    },
    {
      "title": "Andrej Karpathy — official website",
      "url": "https://karpathy.ai/",
      "author": "Andrej Karpathy",
      "published": "Undated; checked 2026-11-13",
      "usage": "Verified the official video links. No assets reproduced.",
      "rights": "Linked reference only.",
      "accessed": "2026-11-13",
      "lessons": []
    }
  ],
  "assets": [
    {
      "asset": "All local SVGs, CSS illustrations, code charts, icons, PNG share artwork, and interactive demonstrations",
      "author": "How LLMs Work project",
      "date": "2026-11-13",
      "rights": "Original project work; no third-party assets reproduced; commercial site use permitted.",
      "modifications": "Created for this release",
      "attribution": "No external attribution required"
    },
    {
      "asset": "Syne variable Latin font",
      "author": "Bonjour Monde, Lucas Descroix, George Triantafyllakos",
      "url": "https://fonts.gstatic.com/s/syne/v24/8vIH7w4qzmVxm2BL9G78HEY.woff2",
      "date": "2026-11-13",
      "license": "SIL Open Font License 1.1",
      "rights": "OFL 1.1 permits use, embedding, bundling and redistribution, including with commercial software; retain copyright and license, do not sell font files by themselves, and comply with reserved-name provisions for modified versions.",
      "attribution": "Copyright 2017 The Syne Project Authors (https://gitlab.com/bonjour-monde/fonderie/syne-typeface)",
      "licensePath": "/assets/2f3627705a248381/fonts/Syne-OFL.txt",
      "file": "/assets/2f3627705a248381/fonts/syne-latin.woff2",
      "modifications": "None by this project. Official CDN WOFF2 bytes retained unchanged. Local descriptive filename only. The Latin subsetting and webfont packaging were supplied by Google Fonts, not created from the PDF.",
      "sha256": "cfff412005eed2f0152d1c110c78ac564642b943e64322fbd687083177d6fa70"
    },
    {
      "asset": "Inter variable Latin font",
      "author": "Rasmus Andersson",
      "url": "https://fonts.gstatic.com/s/inter/v20/UcC73FwrK3iLTeHuS_nVMrMxCp50SjIa1ZL7W0Q5nw.woff2",
      "date": "2026-11-13",
      "license": "SIL Open Font License 1.1",
      "rights": "OFL 1.1 permits use, embedding, bundling and redistribution, including with commercial software; retain copyright and license, do not sell font files by themselves, and comply with reserved-name provisions for modified versions.",
      "attribution": "Copyright 2020 The Inter Project Authors (https://github.com/rsms/inter)",
      "licensePath": "/assets/2f3627705a248381/fonts/Inter-OFL.txt",
      "file": "/assets/2f3627705a248381/fonts/inter-latin.woff2",
      "modifications": "None by this project. Official CDN WOFF2 bytes retained unchanged. Local descriptive filename only. The Latin subsetting and webfont packaging were supplied by Google Fonts, not created from the PDF.",
      "sha256": "c940764593d0fe5d596be327ca7558855e018039fb78509aa21921fd3644c3e4"
    }
  ]
}