Conditional Memory via Scalable Lookup: A New Axis of Sparsity for Large Language Models

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Cheng, Xin, Zeng, Wangding, Dai, Damai, Chen, Qinyu, Wang, Bingxuan, Xie, Zhenda, Huang, Kezhao, Yu, Xingkai, Hao, Zhewen, Li, Yukun, Zhang, Han, Zhang, Huishuai, Zhao, Dongyan, Liang, Wenfeng
Natura: Preprint
Pubblicazione: 2026
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866912817907499008
author Cheng, Xin
Zeng, Wangding
Dai, Damai
Chen, Qinyu
Wang, Bingxuan
Xie, Zhenda
Huang, Kezhao
Yu, Xingkai
Hao, Zhewen
Li, Yukun
Zhang, Han
Zhang, Huishuai
Zhao, Dongyan
Liang, Wenfeng
author_facet Cheng, Xin
Zeng, Wangding
Dai, Damai
Chen, Qinyu
Wang, Bingxuan
Xie, Zhenda
Huang, Kezhao
Yu, Xingkai
Hao, Zhewen
Li, Yukun
Zhang, Han
Zhang, Huishuai
Zhao, Dongyan
Liang, Wenfeng
contents While Mixture-of-Experts (MoE) scales capacity via conditional computation, Transformers lack a native primitive for knowledge lookup, forcing them to inefficiently simulate retrieval through computation. To address this, we introduce conditional memory as a complementary sparsity axis, instantiated via Engram, a module that modernizes classic $N$-gram embedding for O(1) lookup. By formulating the Sparsity Allocation problem, we uncover a U-shaped scaling law that optimizes the trade-off between neural computation (MoE) and static memory (Engram). Guided by this law, we scale Engram to 27B parameters, achieving superior performance over a strictly iso-parameter and iso-FLOPs MoE baseline. Most notably, while the memory module is expected to aid knowledge retrieval (e.g., MMLU +3.4; CMMLU +4.0), we observe even larger gains in general reasoning (e.g., BBH +5.0; ARC-Challenge +3.7) and code/math domains~(HumanEval +3.0; MATH +2.4). Mechanistic analyses reveal that Engram relieves the backbone's early layers from static reconstruction, effectively deepening the network for complex reasoning. Furthermore, by delegating local dependencies to lookups, it frees up attention capacity for global context, substantially boosting long-context retrieval (e.g., Multi-Query NIAH: 84.2 to 97.0). Finally, Engram establishes infrastructure-aware efficiency: its deterministic addressing enables runtime prefetching from host memory, incurring negligible overhead. We envision conditional memory as an indispensable modeling primitive for next-generation sparse models.
format Preprint
id arxiv_https___arxiv_org_abs_2601_07372
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Conditional Memory via Scalable Lookup: A New Axis of Sparsity for Large Language Models
Cheng, Xin
Zeng, Wangding
Dai, Damai
Chen, Qinyu
Wang, Bingxuan
Xie, Zhenda
Huang, Kezhao
Yu, Xingkai
Hao, Zhewen
Li, Yukun
Zhang, Han
Zhang, Huishuai
Zhao, Dongyan
Liang, Wenfeng
Computation and Language
Artificial Intelligence
While Mixture-of-Experts (MoE) scales capacity via conditional computation, Transformers lack a native primitive for knowledge lookup, forcing them to inefficiently simulate retrieval through computation. To address this, we introduce conditional memory as a complementary sparsity axis, instantiated via Engram, a module that modernizes classic $N$-gram embedding for O(1) lookup. By formulating the Sparsity Allocation problem, we uncover a U-shaped scaling law that optimizes the trade-off between neural computation (MoE) and static memory (Engram). Guided by this law, we scale Engram to 27B parameters, achieving superior performance over a strictly iso-parameter and iso-FLOPs MoE baseline. Most notably, while the memory module is expected to aid knowledge retrieval (e.g., MMLU +3.4; CMMLU +4.0), we observe even larger gains in general reasoning (e.g., BBH +5.0; ARC-Challenge +3.7) and code/math domains~(HumanEval +3.0; MATH +2.4). Mechanistic analyses reveal that Engram relieves the backbone's early layers from static reconstruction, effectively deepening the network for complex reasoning. Furthermore, by delegating local dependencies to lookups, it frees up attention capacity for global context, substantially boosting long-context retrieval (e.g., Multi-Query NIAH: 84.2 to 97.0). Finally, Engram establishes infrastructure-aware efficiency: its deterministic addressing enables runtime prefetching from host memory, incurring negligible overhead. We envision conditional memory as an indispensable modeling primitive for next-generation sparse models.
title Conditional Memory via Scalable Lookup: A New Axis of Sparsity for Large Language Models
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2601.07372