[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-84124":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":10,"language":11,"languages":10,"totalLinesOfCode":10,"stars":12,"forks":13,"watchers":14,"openIssues":15,"contributorsCount":16,"subscribersCount":16,"size":16,"stars1d":17,"stars7d":17,"stars30d":17,"stars90d":16,"forks30d":16,"starsTrendScore":18,"compositeScore":19,"rankGlobal":10,"rankLanguage":10,"license":20,"archived":21,"fork":21,"defaultBranch":22,"hasWiki":23,"hasPages":21,"topics":24,"createdAt":10,"pushedAt":44,"updatedAt":45,"readmeContent":46,"aiSummary":10,"trendingCount":16,"starSnapshotCount":16,"syncStatus":17,"lastSyncTime":47,"discoverSource":48},84124,"llm-flashcards","llmsresearch\u002Fllm-flashcards","llmsresearch","300+ visual cards covering almost all large language model(LLMs) concepts and architectures. Best for revising LLM concepts before any big AI\u002FML technical interview rounds.","https:\u002F\u002Fllmsresearch.com\u002Fflashcards",null,"Python",55,4,52,1,0,2,6,47.3,"Other",false,"main",true,[25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43],"agents","ai","anki","attention","deep-learning","fine-tuning","flashcards","gpt","interview-preparation","large-language-models","llm","llm-resources","machine-learning","nlp","prompt-engineering","rag","rlhf","study-notes","transformers","2026-06-08 03:39:22","2026-06-10 04:43:52","# LLM Flashcards\n\nVisual flashcards on how LLMs work.\n\n## The cards\n\n\u003Ctable>\n  \u003Ctr>\n    \u003Ctd width=\"33%\">\u003Ca href=\"cards\u002F01-what-is-a-transformer.jpg\">\u003Cimg src=\"cards\u002F01-what-is-a-transformer.jpg\" alt=\"What is a Transformer?\"\u002F>\u003C\u002Fa>\u003Cp align=\"center\">\u003Cb>Transformer architecture\u003C\u002Fb>\u003C\u002Fp>\u003C\u002Ftd>\n    \u003Ctd width=\"33%\">\u003Ca href=\"cards\u002F02-what-is-tokenization.jpg\">\u003Cimg src=\"cards\u002F02-what-is-tokenization.jpg\" alt=\"What is Tokenization?\"\u002F>\u003C\u002Fa>\u003Cp align=\"center\">\u003Cb>Tokenization\u003C\u002Fb>\u003C\u002Fp>\u003C\u002Ftd>\n    \u003Ctd width=\"33%\">\u003Ca href=\"cards\u002F03-what-is-an-embedding.jpg\">\u003Cimg src=\"cards\u002F03-what-is-an-embedding.jpg\" alt=\"What is an Embedding?\"\u002F>\u003C\u002Fa>\u003Cp align=\"center\">\u003Cb>Embeddings\u003C\u002Fb>\u003C\u002Fp>\u003C\u002Ftd>\n  \u003C\u002Ftr>\n  \u003Ctr>\n    \u003Ctd>\u003Ca href=\"cards\u002F04-language-modeling-objective.jpg\">\u003Cimg src=\"cards\u002F04-language-modeling-objective.jpg\" alt=\"Language Modeling Objective\"\u002F>\u003C\u002Fa>\u003Cp align=\"center\">\u003Cb>Training\u003C\u002Fb>\u003C\u002Fp>\u003C\u002Ftd>\n    \u003Ctd>\u003Ca href=\"cards\u002F05-full-finetuning-vs-peft.jpg\">\u003Cimg src=\"cards\u002F05-full-finetuning-vs-peft.jpg\" alt=\"Full Fine-tuning vs PEFT\"\u002F>\u003C\u002Fa>\u003Cp align=\"center\">\u003Cb>Fine-tuning\u003C\u002Fb>\u003C\u002Fp>\u003C\u002Ftd>\n    \u003Ctd>\u003Ca href=\"cards\u002F06-rlhf-overview.jpg\">\u003Cimg src=\"cards\u002F06-rlhf-overview.jpg\" alt=\"RLHF Overview\"\u002F>\u003C\u002Fa>\u003Cp align=\"center\">\u003Cb>RLHF and alignment\u003C\u002Fb>\u003C\u002Fp>\u003C\u002Ftd>\n  \u003C\u002Ftr>\n  \u003Ctr>\n    \u003Ctd>\u003Ca href=\"cards\u002F07-system-vs-user-prompt.jpg\">\u003Cimg src=\"cards\u002F07-system-vs-user-prompt.jpg\" alt=\"System vs User Prompt\"\u002F>\u003C\u002Fa>\u003Cp align=\"center\">\u003Cb>Prompting\u003C\u002Fb>\u003C\u002Fp>\u003C\u002Ftd>\n    \u003Ctd>\u003Ca href=\"cards\u002F08-what-is-rag.jpg\">\u003Cimg src=\"cards\u002F08-what-is-rag.jpg\" alt=\"What is RAG?\"\u002F>\u003C\u002Fa>\u003Cp align=\"center\">\u003Cb>Retrieval (RAG)\u003C\u002Fb>\u003C\u002Fp>\u003C\u002Ftd>\n    \u003Ctd>\u003Ca href=\"cards\u002F09-what-is-an-llm-agent.jpg\">\u003Cimg src=\"cards\u002F09-what-is-an-llm-agent.jpg\" alt=\"What is an LLM Agent?\"\u002F>\u003C\u002Fa>\u003Cp align=\"center\">\u003Cb>Agents and tools\u003C\u002Fb>\u003C\u002Fp>\u003C\u002Ftd>\n  \u003C\u002Ftr>\n  \u003Ctr>\n    \u003Ctd>\u003Ca href=\"cards\u002F10-autoregressive-generation.jpg\">\u003Cimg src=\"cards\u002F10-autoregressive-generation.jpg\" alt=\"Autoregressive Generation\"\u002F>\u003C\u002Fa>\u003Cp align=\"center\">\u003Cb>Inference\u003C\u002Fb>\u003C\u002Fp>\u003C\u002Ftd>\n    \u003Ctd>\u003Ca href=\"cards\u002F11-scaling-laws.jpg\">\u003Cimg src=\"cards\u002F11-scaling-laws.jpg\" alt=\"Scaling Laws\"\u002F>\u003C\u002Fa>\u003Cp align=\"center\">\u003Cb>Scaling laws\u003C\u002Fb>\u003C\u002Fp>\u003C\u002Ftd>\n    \u003Ctd>\u003Ca href=\"cards\u002F12-gpt-vs-bert-vs-t5.jpg\">\u003Cimg src=\"cards\u002F12-gpt-vs-bert-vs-t5.jpg\" alt=\"GPT vs BERT vs T5\"\u002F>\u003C\u002Fa>\u003Cp align=\"center\">\u003Cb>Architectures\u003C\u002Fb>\u003C\u002Fp>\u003C\u002Ftd>\n  \u003C\u002Ftr>\n  \u003Ctr>\n    \u003Ctd>\u003Ca href=\"cards\u002F13-what-is-quantization.jpg\">\u003Cimg src=\"cards\u002F13-what-is-quantization.jpg\" alt=\"What is Quantization?\"\u002F>\u003C\u002Fa>\u003Cp align=\"center\">\u003Cb>Quantization\u003C\u002Fb>\u003C\u002Fp>\u003C\u002Ftd>\n    \u003Ctd>\u003Ca href=\"cards\u002F14-perplexity-as-a-metric.jpg\">\u003Cimg src=\"cards\u002F14-perplexity-as-a-metric.jpg\" alt=\"Perplexity as a Metric\"\u002F>\u003C\u002Fa>\u003Cp align=\"center\">\u003Cb>Evaluation\u003C\u002Fb>\u003C\u002Fp>\u003C\u002Ftd>\n    \u003Ctd>\u003Ca href=\"cards\u002F15-lost-in-the-middle.jpg\">\u003Cimg src=\"cards\u002F15-lost-in-the-middle.jpg\" alt=\"Lost in the Middle\"\u002F>\u003C\u002Fa>\u003Cp align=\"center\">\u003Cb>Context management\u003C\u002Fb>\u003C\u002Fp>\u003C\u002Ftd>\n  \u003C\u002Ftr>\n  \u003Ctr>\n    \u003Ctd>\u003Ca href=\"cards\u002F16-hallucination.jpg\">\u003Cimg src=\"cards\u002F16-hallucination.jpg\" alt=\"Hallucination\"\u002F>\u003C\u002Fa>\u003Cp align=\"center\">\u003Cb>Safety and ethics\u003C\u002Fb>\u003C\u002Fp>\u003C\u002Ftd>\n    \u003Ctd>\u003Ca href=\"cards\u002F17-chat-completion-api.jpg\">\u003Cimg src=\"cards\u002F17-chat-completion-api.jpg\" alt=\"Chat Completion API\"\u002F>\u003C\u002Fa>\u003Cp align=\"center\">\u003Cb>APIs and practical\u003C\u002Fb>\u003C\u002Fp>\u003C\u002Ftd>\n    \u003Ctd>\u003Ca href=\"cards\u002F18-multimodal-llms.jpg\">\u003Cimg src=\"cards\u002F18-multimodal-llms.jpg\" alt=\"Multimodal LLMs\"\u002F>\u003C\u002Fa>\u003Cp align=\"center\">\u003Cb>Multimodal\u003C\u002Fb>\u003C\u002Fp>\u003C\u002Ftd>\n  \u003C\u002Ftr>\n  \u003Ctr>\n    \u003Ctd>\u003Ca href=\"cards\u002F19-reasoning-in-llms.jpg\">\u003Cimg src=\"cards\u002F19-reasoning-in-llms.jpg\" alt=\"Reasoning in LLMs\"\u002F>\u003C\u002Fa>\u003Cp align=\"center\">\u003Cb>Reasoning\u003C\u002Fb>\u003C\u002Fp>\u003C\u002Ftd>\n    \u003Ctd>\u003Ca href=\"cards\u002F20-reasoning-models.jpg\">\u003Cimg src=\"cards\u002F20-reasoning-models.jpg\" alt=\"Reasoning Models\"\u002F>\u003C\u002Fa>\u003Cp align=\"center\">\u003Cb>Reasoning models\u003C\u002Fb>\u003C\u002Fp>\u003C\u002Ftd>\n    \u003Ctd>\u003Ca href=\"cards\u002F21-state-space-models-mamba.jpg\">\u003Cimg src=\"cards\u002F21-state-space-models-mamba.jpg\" alt=\"State Space Models and Mamba\"\u002F>\u003C\u002Fa>\u003Cp align=\"center\">\u003Cb>Architectures\u003C\u002Fb>\u003C\u002Fp>\u003C\u002Ftd>\n  \u003C\u002Ftr>\n  \u003Ctr>\n    \u003Ctd>\u003Ca href=\"cards\u002F22-mixture-of-experts-routing.jpg\">\u003Cimg src=\"cards\u002F22-mixture-of-experts-routing.jpg\" alt=\"Mixture of Experts Routing\"\u002F>\u003C\u002Fa>\u003Cp align=\"center\">\u003Cb>Architectures\u003C\u002Fb>\u003C\u002Fp>\u003C\u002Ftd>\n    \u003Ctd>\u003Ca href=\"cards\u002F23-model-context-protocol.jpg\">\u003Cimg src=\"cards\u002F23-model-context-protocol.jpg\" alt=\"Model Context Protocol\"\u002F>\u003C\u002Fa>\u003Cp align=\"center\">\u003Cb>Agents and tools\u003C\u002Fb>\u003C\u002Fp>\u003C\u002Ftd>\n    \u003Ctd>\u003Ca href=\"cards\u002F24-vision-transformer.jpg\">\u003Cimg src=\"cards\u002F24-vision-transformer.jpg\" alt=\"Vision Transformer\"\u002F>\u003C\u002Fa>\u003Cp align=\"center\">\u003Cb>Multimodal\u003C\u002Fb>\u003C\u002Fp>\u003C\u002Ftd>\n  \u003C\u002Ftr>\n  \u003Ctr>\n    \u003Ctd>\u003Ca href=\"cards\u002F25-sparse-autoencoders.jpg\">\u003Cimg src=\"cards\u002F25-sparse-autoencoders.jpg\" alt=\"Sparse Autoencoders\"\u002F>\u003C\u002Fa>\u003Cp align=\"center\">\u003Cb>Interpretability\u003C\u002Fb>\u003C\u002Fp>\u003C\u002Ftd>\n    \u003Ctd>\u003Ca href=\"cards\u002F26-tree-of-thoughts.jpg\">\u003Cimg src=\"cards\u002F26-tree-of-thoughts.jpg\" alt=\"Tree of Thoughts\"\u002F>\u003C\u002Fa>\u003Cp align=\"center\">\u003Cb>Prompting\u003C\u002Fb>\u003C\u002Fp>\u003C\u002Ftd>\n    \u003Ctd>\u003Ca href=\"cards\u002F27-double-descent.jpg\">\u003Cimg src=\"cards\u002F27-double-descent.jpg\" alt=\"Double Descent\"\u002F>\u003C\u002Fa>\u003Cp align=\"center\">\u003Cb>Training\u003C\u002Fb>\u003C\u002Fp>\u003C\u002Ftd>\n  \u003C\u002Ftr>\n  \u003Ctr>\n    \u003Ctd>\u003Ca href=\"cards\u002F28-activation-functions.jpg\">\u003Cimg src=\"cards\u002F28-activation-functions.jpg\" alt=\"Activation Functions\"\u002F>\u003C\u002Fa>\u003Cp align=\"center\">\u003Cb>Training\u003C\u002Fb>\u003C\u002Fp>\u003C\u002Ftd>\n    \u003Ctd>\u003Ca href=\"cards\u002F29-gpqa.jpg\">\u003Cimg src=\"cards\u002F29-gpqa.jpg\" alt=\"GPQA\"\u002F>\u003C\u002Fa>\u003Cp align=\"center\">\u003Cb>Evaluation\u003C\u002Fb>\u003C\u002Fp>\u003C\u002Ftd>\n    \u003Ctd>\u003Ca href=\"cards\u002F30-matryoshka-embeddings.jpg\">\u003Cimg src=\"cards\u002F30-matryoshka-embeddings.jpg\" alt=\"Matryoshka Embeddings\"\u002F>\u003C\u002Fa>\u003Cp align=\"center\">\u003Cb>Embeddings\u003C\u002Fb>\u003C\u002Fp>\u003C\u002Ftd>\n  \u003C\u002Ftr>\n\u003C\u002Ftable>\n\nClick any card to open it full size.\n\n**Study in Anki:** download [`llm-flashcards.apkg`](llm-flashcards.apkg) (these 30 cards) and import it into [Anki](https:\u002F\u002Fapps.ankiweb.net\u002F). Front is the concept, back is the card.\n\n## Why I made them\n\nI work on LLM efficiency at LLMs Research, and a lot of that work happens on a whiteboard. Drawing a thing forces you to know what you're drawing. A vague hand-wave on a slide hides confusion. A diagram doesn't.\n\nAfter enough whiteboards I had a stack of diagrams. The stack turned into a study set for myself. I tightened the lines, kept the labels honest, and put them on cards. That's the set.\n\nThe cards are for someone who has used an LLM API and wants the layer underneath. Some technical background helps. No heavy math.\n\n## What's in the full set\n\n332 cards across 22 topics:\n\n| | | |\n|---|---|---|\n| Tokenization (12) | Embeddings and retrieval (14) | Transformer architecture (30) |\n| Architecture variants (16) | Training (18) | Distributed training (10) |\n| Scaling laws (10) | Fine-tuning (15) | RLHF and alignment (19) |\n| Inference and decoding (19) | Quantization (12) | Prompting (19) |\n| Reasoning (15) | Context management (10) | RAG (24) |\n| Agents and tools (22) | Multimodal (8) | Advanced concepts (6) |\n| Evaluation (16) | Safety (17) | Interpretability (7) |\n| APIs and practical use (13) | | |\n\nThree formats: a PDF (332 pages, printable), an `.apkg` for Anki spaced-repetition review, and every card as a separate image. New cards get added regularly, and past buyers get every update free.\n\n[llmsresearch.com\u002Fflashcards](https:\u002F\u002Fllmsresearch.com\u002Fflashcards?utm_source=github&utm_medium=repo&utm_campaign=flashcards_launch)\n\n## License\n\nCC BY-NC-ND 4.0. Share the cards with credit and a link back to this repo. No repackaging, no reselling, no modified versions, no commercial use. Full text in [LICENSE](LICENSE).\n\n## Contributing\n\nIf something on a card is wrong or unclear, [open an issue](..\u002F..\u002Fissues\u002Fnew). If you want a card on a concept that is not in the set yet, open one too. I read them.\n\n## About\n\n[LLMs Research](https:\u002F\u002Fllmsresearch.com) is an independent applied research lab. We work on LLM efficiency: inference, KV cache compression, adaptive compute, multi-agent systems. The set started as study notes for that work.\n\n[Website](https:\u002F\u002Fllmsresearch.com) · [Newsletter](https:\u002F\u002Fllmsresearch.substack.com) · [X](https:\u002F\u002Fx.com\u002Fllmsresearch) · [LinkedIn](https:\u002F\u002Fwww.linkedin.com\u002Fcompany\u002Fllmsresearch)\n","2026-06-11 04:12:20","CREATED_QUERY"]