https://llm.jerrinot.info/https://llm.jerrinot.info/architectures/gemma4-overview/https://llm.jerrinot.info/architectures/gqa-and-mqa/https://llm.jerrinot.info/architectures/moe-aggregation/https://llm.jerrinot.info/architectures/moe-router/https://llm.jerrinot.info/architectures/shared-kv-layers/https://llm.jerrinot.info/architectures/sliding-window-attention/https://llm.jerrinot.info/block/attention-output/https://llm.jerrinot.info/block/causal-attention/https://llm.jerrinot.info/block/feed-forward-network/https://llm.jerrinot.info/block/full-decoder-block/https://llm.jerrinot.info/block/multi-head-attention/https://llm.jerrinot.info/block/position-and-rope/https://llm.jerrinot.info/block/qkv-projections/https://llm.jerrinot.info/block/residuals-and-rmsnorm/https://llm.jerrinot.info/block/what-a-layer-does/https://llm.jerrinot.info/capstone/zero-to-hero/https://llm.jerrinot.info/case-study/dense-gemma/https://llm.jerrinot.info/case-study/gemma4-attention/https://llm.jerrinot.info/case-study/gemma4-moe/https://llm.jerrinot.info/case-study/optimization-and-regression/https://llm.jerrinot.info/case-study/profile-interpretation/https://llm.jerrinot.info/dashboard/progress/https://llm.jerrinot.info/inference/batching-and-ubatching/https://llm.jerrinot.info/inference/context-scaling/https://llm.jerrinot.info/inference/continuous-batching/https://llm.jerrinot.info/inference/decode/https://llm.jerrinot.info/inference/kv-cache/https://llm.jerrinot.info/inference/prefill/https://llm.jerrinot.info/math/dot-product/https://llm.jerrinot.info/math/embeddings-as-vectors/https://llm.jerrinot.info/math/linear-algebra-synthesis/https://llm.jerrinot.info/math/linear-projections/https://llm.jerrinot.info/math/matrix-multiply/https://llm.jerrinot.info/math/scalars-vectors/https://llm.jerrinot.info/math/tensor-shapes/https://llm.jerrinot.info/performance/compute-vs-memory/https://llm.jerrinot.info/performance/gemm-vs-gemv/https://llm.jerrinot.info/performance/quantization/https://llm.jerrinot.info/performance/repack-and-layout/https://llm.jerrinot.info/performance/threads-and-affinity/https://llm.jerrinot.info/performance/validate-speedups/https://llm.jerrinot.info/performance/where-time-goes/https://llm.jerrinot.info/probability/argmax-and-sampling/https://llm.jerrinot.info/probability/foundations-checkpoint/https://llm.jerrinot.info/probability/logits/https://llm.jerrinot.info/probability/softmax/https://llm.jerrinot.info/quiz/architecture-variants/https://llm.jerrinot.info/quiz/dense-transformer/https://llm.jerrinot.info/quiz/inference-mechanics/https://llm.jerrinot.info/quiz/performance-reasoning/https://llm.jerrinot.info/reference/glossary/https://llm.jerrinot.info/reference/live-inference/https://llm.jerrinot.info/start/one-request-end-to-end/https://llm.jerrinot.info/start/what-an-llm-does/https://llm.jerrinot.info/start/what-you-will-learn/https://llm.jerrinot.info/tokens/context-length/https://llm.jerrinot.info/tokens/discrete-input-synthesis/https://llm.jerrinot.info/tokens/vocabulary-and-ids/https://llm.jerrinot.info/tokens/what-is-a-token/