pith. sign in

Title resolution pending

1 Pith paper cite this work. Polarity classification is still indexing.

1 Pith paper citing it

fields

cs.LG 1

years

2025 1

verdicts

UNVERDICTED 1

representative citing papers

The Bayesian Geometry of Transformer Attention

cs.LG · 2025-12-27 · unverdicted · novelty 7.0

Small transformers reproduce known Bayesian posteriors with 10^{-3} to 10^{-4} bit accuracy in verifiable wind-tunnel tasks via residual belief states, FFN updates, and attention routing, while MLPs do not.

citing papers explorer

Showing 1 of 1 citing paper.

  • The Bayesian Geometry of Transformer Attention cs.LG · 2025-12-27 · unverdicted · none · ref 14

    Small transformers reproduce known Bayesian posteriors with 10^{-3} to 10^{-4} bit accuracy in verifiable wind-tunnel tasks via residual belief states, FFN updates, and attention routing, while MLPs do not.