Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

An Integrable Token Mixing Layer from the Generalized Yang Baxter Equation

About

The YB Mixer is a sequence token mixing layer derived from free fermion and generalized Yang Baxter structures. It applies a core principle from integrable systems where a local algebraic constraint guarantees global computational stability. By using the Ising exchange algebra the mixer creates a free fermionic structure that acts as an exactly norm preserving orthogonal map. This algebra also produces commuting transfer matrices which allow inference to be order free and adaptable to any variable budget. To ensure the model can generalize to longer sequence lengths it uses a spectral circulant generator. This generator maintains the crucial orthogonal and commuting properties of the system. The result is a highly stable and mathematically grounded architecture for sequence processing.

Snigdha Chandan Khilar• 2026

Related benchmarks

TaskDatasetResultRank
Memory taskLong-range memory L=24 (test)
Accuracy100
6
Recall taskContent-dependent recall L=17 (test)
Accuracy75.9
6
Image ClassificationMNIST permuted (val)
Accuracy98.6
5
Image ClassificationLRA-Image seq-CIFAR (val)
Accuracy84.8
5
Token RetrievalInduction Heads (val)
Accuracy100
5
List OperationsLRA-ListOps (val)
Accuracy33.1
5
Text ClassificationLRA-Text byte-IMDB (val)
Accuracy79.4
5
Showing 7 of 7 rows

Other info

Follow for update