Model Architecture & Training MethodsGreg Yang et al.

muP / Tensor Programs V

Made by AIWritten by AI (Claude) from the source below. Check anything important against the source.How it's made

muP / Tensor Programs V

Why it matters

Transferred hyperparameters tuned on small models zero-shot to large ones, making the practice of getting it right in a single run possible

Source — Primary sources first; secondary reports are labelled “Press”

Paperarxiv.orghttps://arxiv.org/abs/2203.03466

Added in Week 35, 2026 · See that week

On the map — ⑪ Model Architecture line

  1. Previous stopSwitch Transformer
  2. This recordmuP / Tensor Programs V
  3. Next stopChinchilla
See it on the map

Same week

Model SoupsPaperMarcus writes Deep Learning Is Hitting a Wall for NautilusOfficial

Nearby on the same axis

Model SoupsPaperChinchillaPaperGoogle Research announces PaLM (540B)OfficialEmergent Abilities of LLMsPaper