Multi-Head Attention Scaling and Per-Head Gain
SeunghyunSEO7 · x · 2026-07-16
During a discussion on multi-head attention, someone pointed out that many LLMs currently apply qnorm directly to qproj(x), rather than scaling each head's q/k by \sqrt(dhead).
Replies mentioned that the 1/dhead scaling concept originates from muP. The reason for following a Law of Large Numbers (LLN) intuition is that q and k might already be aligned from the start. Another viewpoint suggested that if q and k are first scaled by a factor like \sqrt{dhead} gain, their dot product becomes dhead (qgain, kgain) cos(θ), meaning the "per head gain" design warrants further scrutiny.
Related event: Researchers Debate 1/d-Style Attention Scaling in Modern LLMs(6 posts)→
More from Research
- 3D ResNet Paper Crosses 3,000 Citations Eight Years After CVPR 2018 — HirokatuKataoka · 2026-09-11
- Sample selection and ordering matter a lot in LLM training: DataFlex makes data scheduling dynamic — Puzzleheaded_Box2842 · 2026-09-11
- Jeff Heaton's Intro to the Math of Neural Networks eBook Is Free to Download — blaizedsouza · 2026-09-11
- Mathematician Daniel Litt Launches Problem Repo to Track Human vs AI Progress: 15 Problems, 1 Solved — littmath · 2026-09-11
- Open ECDSA.fail challenge uses AI agents to shrink Shor's-algorithm quantum circuits for Bitcoin keys — StefanoGogioso · 2026-09-11
- Alex Townsend posts 200 open problems in numerical linear algebra for humans and AI agents — IgorCarron · 2026-09-11