{
  "slug": "transformer-rank-collapse-how-skips-norms-and-mlp-width-affect-trainability-714018",
  "title": "Transformer Rank Collapse: How Skips, Norms, and MLP Width Affect Trainability",
  "dek": "Katie Everett’s paper argues that residual skips, normalization placement, and feedforward expansion help determine how much gradient rank survives as Transformer depth increases.",
  "summary": "How Transformer rank collapse is shaped by residual skips, Pre-Norm vs Post-Norm, and feedforward expansion—and why rank at initialization may predict trainability.",
  "tags": [
    "transformers",
    "rank collapse",
    "deep learning",
    "normalization",
    "residual connections",
    "random matrix theory"
  ],
  "published_at": "2026-07-16T14:03:03.614+00:00",
  "grade": 8,
  "agent_utility": null,
  "price_usdc": 0.01,
  "stats": {
    "claims": 0,
    "entities": 0,
    "faq": 0,
    "downloads": 0,
    "paid_downloads": 0
  },
  "preview": {
    "faq_questions": [],
    "entity_names": [],
    "related_work_titles": [],
    "application_industries": [],
    "glossary_terms": []
  },
  "purchase": {
    "protocol": "x402",
    "version": 2,
    "network": "eip155:8453",
    "asset": "USDC",
    "price": 0.01,
    "full_markdown_url": "https://x402.disruptive-concepts.com/api/public/article/transformer-rank-collapse-how-skips-norms-and-mlp-width-affect-trainability-714018/full",
    "full_json_url": "https://x402.disruptive-concepts.com/api/public/article/transformer-rank-collapse-how-skips-norms-and-mlp-width-affect-trainability-714018/full?format=json"
  }
}