Salvato in:
Dettagli Bibliografici
Autori principali: Chen, Yilong, Xie, Yanxi, Gao, Zitian, Xin, He, Xiao, Yihao, Liu, Jason Klein, Luo, Haoming, Luo, Yifan, Ye, Zhengmao, Liu, Tingwen, Zhao, Xin, Tao, Ran, Dai, Bryan
Natura: Preprint
Pubblicazione: 2026
Soggetti:
Accesso online:https://arxiv.org/abs/2604.21724
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866910161786896384
author Chen, Yilong
Xie, Yanxi
Gao, Zitian
Xin, He
Xiao, Yihao
Liu, Jason Klein
Luo, Haoming
Luo, Yifan
Ye, Zhengmao
Liu, Tingwen
Zhao, Xin
Tao, Ran
Dai, Bryan
author_facet Chen, Yilong
Xie, Yanxi
Gao, Zitian
Xin, He
Xiao, Yihao
Liu, Jason Klein
Luo, Haoming
Luo, Yifan
Ye, Zhengmao
Liu, Tingwen
Zhao, Xin
Tao, Ran
Dai, Bryan
contents Large token-indexed lookup tables provide a compute-decoupled scaling path, but their practical gains are often limited by poor parameter efficiency and rapid memory growth. We attribute these limitations to Zipfian under-training of the long tail, heterogeneous demand across layers, and "slot collapse" that produces redundant embeddings. To address this, we propose X-GRAM, a frequency-aware dynamic token-injection framework. X-GRAM employs hybrid hashing and alias mixing to compress the tail while preserving head capacity, and refines retrieved vectors via normalized SwiGLU ShortConv to extract diverse local n-gram features. These signals are integrated into attention value streams and inter-layer residuals using depth-aware gating, effectively aligning static memory with dynamic context. This design introduces a memory-centric scaling axis that decouples model capacity from FLOPs. Extensive evaluations at the 0.73B and 1.15B scales show that X-GRAM improves average accuracy by as much as 4.4 points over the vanilla backbone and 3.2 points over strong retrieval baselines, while using substantially smaller tables in the 50% configuration. Overall, by decoupling capacity from compute through efficient memory management, X-GRAM offers a scalable and practical paradigm for future memory-augmented architectures. Code aviliable in https://github.com/Longyichen/X-gram.
format Preprint
id arxiv_https___arxiv_org_abs_2604_21724
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Beyond N-gram: Data-Aware X-GRAM Extraction for Efficient Embedding Parameter Scaling
Chen, Yilong
Xie, Yanxi
Gao, Zitian
Xin, He
Xiao, Yihao
Liu, Jason Klein
Luo, Haoming
Luo, Yifan
Ye, Zhengmao
Liu, Tingwen
Zhao, Xin
Tao, Ran
Dai, Bryan
Computation and Language
Large token-indexed lookup tables provide a compute-decoupled scaling path, but their practical gains are often limited by poor parameter efficiency and rapid memory growth. We attribute these limitations to Zipfian under-training of the long tail, heterogeneous demand across layers, and "slot collapse" that produces redundant embeddings. To address this, we propose X-GRAM, a frequency-aware dynamic token-injection framework. X-GRAM employs hybrid hashing and alias mixing to compress the tail while preserving head capacity, and refines retrieved vectors via normalized SwiGLU ShortConv to extract diverse local n-gram features. These signals are integrated into attention value streams and inter-layer residuals using depth-aware gating, effectively aligning static memory with dynamic context. This design introduces a memory-centric scaling axis that decouples model capacity from FLOPs. Extensive evaluations at the 0.73B and 1.15B scales show that X-GRAM improves average accuracy by as much as 4.4 points over the vanilla backbone and 3.2 points over strong retrieval baselines, while using substantially smaller tables in the 50% configuration. Overall, by decoupling capacity from compute through efficient memory management, X-GRAM offers a scalable and practical paradigm for future memory-augmented architectures. Code aviliable in https://github.com/Longyichen/X-gram.
title Beyond N-gram: Data-Aware X-GRAM Extraction for Efficient Embedding Parameter Scaling
topic Computation and Language
url https://arxiv.org/abs/2604.21724