@misc{gong2026express,
author = {Gong, Albert and Carrell, A. and Dwivedi, Raaz and Mackey, Lester},
title = {Express Language Modeling},
howpublished = {arXiv},
year = {2026},
month = {June},
abstract = {We introduce a new tool, Express, for converting a non-causal attention approximation into a causal approximation with matching approximation guarantees. When combined with the state-of-the-art Thinformer approximation, Express improves upon the best known causal attention guarantees, delivering log3/2(n)/s approximation error with only O(s) memory and O(s2log2(n)) compression overhead for a sequence of length n. We pair these developments with an efficient I/O-aware Triton implementation, demonstrate substantial speedups over FlashAttention 2, and use Express to overcome four resource bottlenecks in the language modeling pipeline: long-context prefill, KV cache compression, long-form memory-constrained decoding, and long-form compute-constrained decoding.},
url = {http://approjects.co.za/?big=en-us/research/publication/express-language-modeling/},
}