's   

Rushi's

Ctrl+AI+Ship

  • Home
  • Musings
  • Tech
  • About
  • Contact

Day: June 2, 2026

Jun 02
2026
0

LLM parameters: what they are and how they actually work

Posted by Rushi

You’ve seen the numbers. 7B. 70B. 405B. Everyone talks about parameter counts. But what are they? Why does size matter? And what actually happens when you hit “Generate”? This post covers the mechanics: what parameters are, where they live in the model architecture, how scaling affects them, and what that means if you’re running or choosing […]

Read More →
tech a100, activation-functions, ai-engineering, artificial-intelligence, attention-mechanism, beginner-friendly, BF16, biases, chinchilla, compute-optimal, deep-learning, embedding-layer, expert-routing, explainer, feed-forward-network, foundation-models, FP16, generative-ai, GGUF, gpu-memory, Grok, inference, inference-optimization, int4, int8, kv-cache, large-language-models, Llama, llama-3, llama-cpp, llm, local-llm, machine-learning, mixtral, mixtral-8x7b, mixture-of-experts, ml-engineering, mlops, mlp, model-compression, model-parameters, model-serving, model-size, moe, neural-networks, neurons, nlp, on-premise-ai, Parameters, pretraining, quantization, query-key-value, relu, scaling-laws, self-attention, self-hosted-ai, sparse-models, swiglu, technical-deep-dive, training, transformer, vram, weights

Tags

ai AI agents AI coding agents angularjs anthropic artificial intelligence automation browser Chrome claude Claude Code code css cursor design developer tools git Google html images java javascript js linux llm LLMs machine learning MCP nasa node.js ollama open source pics productivity programming prompt engineering Python Research software engineering Spec-Driven Development typescript video videos Windows youtube

RSS RSS

  • From Prompting Agents to Loop Engineering
  • Local RAG Without the Storage Tax: A Hands-On Guide to LEANN
  • Building Your First Agent with Flue
  • The AI native playbook: people, agents, and context
  • Scanning agent skills before you trust them: a look at NVIDIA SkillSpector
June 2026
MTWTFSS
1234567
891011121314
15161718192021
22232425262728
2930 
« May    
© 2026  rushis.com. | The content is copyrighted to Rushi and may not be reproduced.