https://mosthumble.github.io/posts/Recall-and-Precision-A-Practical-Case-Against-Memorization/ 2025-11-16T00:37:05+01:00 https://mosthumble.github.io/posts/Teaching-an-AI-to-Drive-a-Taxi-A-Friendly-Guide-to-Q-Learning/ 2025-02-11T15:02:55+01:00 https://mosthumble.github.io/posts/Vanishing-and-exploding-Gradients-A-non-flat-earther's-perspective/ 2025-05-24T19:06:58+02:00 https://mosthumble.github.io/posts/The-Hidden-Beauty-of-Sinusoidal-Positional-Encodings-in-Transformers/ 2025-02-11T15:02:19+01:00 https://mosthumble.github.io/posts/Attention-scores,-Scaling-and-Softmax/ 2026-02-08T11:21:35+01:00 https://mosthumble.github.io/posts/KV-cache-The-how-not-to-waste-your-FLOPS-starter/ 2025-06-01T10:47:32+02:00 https://mosthumble.github.io/posts/A-beginner's-guide-to-Vision-Language-Models-(VLMs)/ 2026-02-08T11:24:11+01:00 https://mosthumble.github.io/posts/Guidance-Structuring-your-outputs-is-easier-than-you-think/ 2026-02-08T11:24:11+01:00 https://mosthumble.github.io/posts/Doing-MORE-To-consume-LESS-Flash-Attention-V1/ 2025-02-11T15:02:19+01:00 https://mosthumble.github.io/posts/Mixture-of-Experts-Scaling-Transformers-Without-Breaking-the-FLOPS-Bank/ 2026-02-08T11:24:11+01:00 https://mosthumble.github.io/posts/Speculative-Decoding-Making-Language-Models-Generate-Faster-Without-Losing-Their-Minds/ 2026-02-08T11:24:11+01:00 https://mosthumble.github.io/posts/Chunking-Strategies-for-RAG-Breaking-Down-Documents-for-Better-Retrieval/ 2026-02-08T11:24:11+01:00 https://mosthumble.github.io/posts/Do-You-Need-A-Matryoshka-Model/ 2025-07-19T11:04:41+02:00 https://mosthumble.github.io/posts/SFT-vs-DPO-A-Visual-Guide-to-LLM-Fine-Tuning/ 2025-10-23T00:28:03+02:00 https://mosthumble.github.io/posts/Elements-Of-Mechanistic-Interpretability-From-Observation-to-Causation/ 2025-10-29T22:59:39+01:00 https://mosthumble.github.io/posts/Entropic-Instruction-Following/ 2025-12-07T18:41:42+01:00 https://mosthumble.github.io/posts/The-Epsilon-Trap-When-Adam-Stops-Being-Adam/ 2026-02-17T20:56:01+01:00 https://mosthumble.github.io/posts/Why-Modern-LLMs-Dropped-Mean-Centering-(And-Got-Away-With-It)/ 2026-02-23T09:52:20+01:00 https://mosthumble.github.io/posts/Towards-a-Bitter-Lesson-of-Optimization-When-Neural-Networks-Write-Their-Own-Update-Rules/ 2026-04-24T22:25:45+02:00 https://mosthumble.github.io/categories/ 2026-04-24T22:27:46+02:00 https://mosthumble.github.io/tags/ 2026-04-24T22:27:46+02:00 https://mosthumble.github.io/archives/ 2026-04-24T22:27:46+02:00 https://mosthumble.github.io/about/ 2026-04-24T22:27:46+02:00 https://mosthumble.github.io/ https://mosthumble.github.io/tags/metrics/ https://mosthumble.github.io/tags/interview/ https://mosthumble.github.io/tags/q-learning/ https://mosthumble.github.io/tags/gymnasium/ https://mosthumble.github.io/tags/python/ https://mosthumble.github.io/tags/tutorial/ https://mosthumble.github.io/tags/normalization/ https://mosthumble.github.io/tags/gradients/ https://mosthumble.github.io/tags/math/ https://mosthumble.github.io/tags/positional-embedding/ https://mosthumble.github.io/tags/transformers/ https://mosthumble.github.io/tags/training-stability/ https://mosthumble.github.io/tags/convergence/ https://mosthumble.github.io/tags/memory/ https://mosthumble.github.io/tags/inference/ https://mosthumble.github.io/tags/optimizations/ https://mosthumble.github.io/tags/flops/ https://mosthumble.github.io/tags/contrastive-learning/ https://mosthumble.github.io/tags/vision/ https://mosthumble.github.io/tags/grammar/ https://mosthumble.github.io/tags/structured-generation/ https://mosthumble.github.io/tags/computational-efficiency/ https://mosthumble.github.io/tags/guidance/ https://mosthumble.github.io/tags/hardware/ https://mosthumble.github.io/tags/efficiency/ https://mosthumble.github.io/tags/scaling/ https://mosthumble.github.io/tags/moe/ https://mosthumble.github.io/tags/gating/ https://mosthumble.github.io/tags/llms/ https://mosthumble.github.io/tags/sampling/ https://mosthumble.github.io/tags/speculative-decoding/ https://mosthumble.github.io/tags/nlp/ https://mosthumble.github.io/tags/rag/ https://mosthumble.github.io/tags/chunking/ https://mosthumble.github.io/tags/embeddings/ https://mosthumble.github.io/tags/retrieval/ https://mosthumble.github.io/tags/indexing/ https://mosthumble.github.io/tags/llm/ https://mosthumble.github.io/tags/fine-tuning/ https://mosthumble.github.io/tags/sft/ https://mosthumble.github.io/tags/dpo/ https://mosthumble.github.io/tags/rlhf/ https://mosthumble.github.io/tags/interpretability/ https://mosthumble.github.io/tags/mechanistic-interpretability/ https://mosthumble.github.io/tags/transformers/ https://mosthumble.github.io/tags/probing/ https://mosthumble.github.io/tags/activation-patching/ https://mosthumble.github.io/tags/evaluation/ https://mosthumble.github.io/tags/instruction-following/ https://mosthumble.github.io/tags/context-window/ https://mosthumble.github.io/tags/optimization/ https://mosthumble.github.io/tags/adam/ https://mosthumble.github.io/tags/pytorch/ https://mosthumble.github.io/tags/normalization/ https://mosthumble.github.io/tags/geometry/ https://mosthumble.github.io/tags/meta-learning/ https://mosthumble.github.io/categories/machine-learning/ https://mosthumble.github.io/categories/evaluation/ https://mosthumble.github.io/categories/reinforcement-learning/ https://mosthumble.github.io/categories/deep-learning/ https://mosthumble.github.io/categories/vision-language-models/ https://mosthumble.github.io/categories/language-models/ https://mosthumble.github.io/categories/optimization/ https://mosthumble.github.io/categories/adam/ https://mosthumble.github.io/categories/meta-learning/ https://mosthumble.github.io/page2/