RadixMLP - Intra-batch Deduplication for Causal Transformers
Michael Feil, Julius Lipp · 2026
Read RadixMLP - Intra-batch Deduplication for Causal Transformers by Michael Feil, Julius Lipp (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-15013.txt.
Description
Michael Feil and Julius Lipp speed up a repeated step inside transformer models. Their 2026 arXiv paper deals with batch inference, where many text sequences run through the model at once. Those sequences often share a common prefix, such as a system prompt.
Standard engines treat each sequence on its own. They recompute the same MLP activations for each copy of the shared prefix. The authors present RadixMLP, which deduplicates that work inside the batch. Shared prefix work is done once, and the savings in time and compute grow with the batch.
Catalog record
| CREATOR | Michael Feil, Julius Lipp |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 60 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2601-15013