{
 "slug": "RMSNorm",
 "title": "RMSNorm",
 "type": "concept",
 "short_desc": "Root mean square layer normalization: a simplification of layer normalization that rescales activations by their root mean square alone, the default normalizer of modern open-weight LLMs.",
 "categories": [
  "Architectures",
  "Concepts"
 ],
 "infobox": {
  "Introduced": "Zhang and Sennrich, 2019",
  "Field": "Neural network architectures, normalization",
  "Used in": "LLaMA, Gemma, Qwen, DeepSeek, and most open-weight models since 2023",
  "Key papers": "arXiv:1910.07467, arXiv:2302.13971"
 },
 "infobox_links": {
  "Used in": [
   "LLaMA",
   "Gemma",
   "Qwen_(model_family)",
   "DeepSeek_(model_family)"
  ]
 },
 "aliases": [
  "Root mean square layer normalization"
 ],
 "family": "",
 "org": "",
 "registry_status": "stub",
 "words": 1149,
 "references": 5,
 "outbound": [
  "BERT",
  "DeepSeek_(model_family)",
  "Feed-forward_network",
  "Gemma",
  "Google_DeepMind",
  "Gopher",
  "LLaMA",
  "Large_language_model",
  "Layer_normalization",
  "Meta_AI",
  "Mistral_7B",
  "Qwen_(model_family)",
  "Self-attention",
  "T5",
  "Transformer_(architecture)"
 ],
 "inbound": [
  "DeepSeek_LLM",
  "Feed-forward_network",
  "Gopher",
  "LLaMA",
  "Layer_normalization",
  "Qwen_1",
  "The_Illustrated_Transformer",
  "Transformer_(architecture)"
 ],
 "url": "wiki/RMSNorm.html",
 "built": "2026-07-24"
}
