| title | Long Context Pre-Training with Lighthouse Attention |
|---|---|
| authors | Bowen Peng, Subho Ghosh, Jeffrey Quesnelle |
| date | 2026-05-07 |
| venue | arXiv (preprint, not peer reviewed) |
| id | arXiv:2605.06554 |
| affiliation | NousResearch |
| upvotes | 30 (Hugging Face Daily Papers, 2026-11-01) |
| links | abs pdf hf code |
| abstract | Training causal transformers at extreme sequence lengths is bottlenecked by the quadratic time and memory of scaled dot-product attention (SDPA). In this work, we propose Lighthouse Attention, a training-only symmetrical selection-based hierarchical attention algorithm that wraps around ordinary SDPA and can be easily removed towards the end of the training. Our hierarchical selection is also gradient-free, which exempts us from dealing with a complicated and potentially inefficient backward pass kernel. Our contribution is three-fold: (i) A subquadratic hierarchical pre- and post-processing step that does adaptive compression and decompression of the sequence. (ii) A symmetrical compression strategy that pools queries, keys and values at the same time, while preserving left-to-right causality, which greatly improves parallelism. (iii) A two stage training approach which we pre-train for the majority of the time with Lighthouse Attention and recover a full attention model at the end with a short training. We run preliminary small scale LLM pre-training experiments that show the effectiveness of our method compared to full attention training with all other settings matched, where we achieve a faster total training time and lower final loss after the recovery phase. Full code is available at: https://github.com/ighoshsubho/lighthouse-attention |
cite
@misc{peng2026long,
title={Long Context Pre-Training with Lighthouse Attention},
author={Bowen Peng and Subho Ghosh and Jeffrey Quesnelle},
year={2026},
eprint={2605.06554},
archivePrefix={arXiv},
url={https://arxiv.org/abs/2605.06554}
}