CORTEXA
← Browse
arxivcs.CLcs.AIcs.LG2026-06-29

MATCH: Modulating Attention via In-Context Retrieval for Long-Context Transformers

Linrui Ma, Chun Hei Lo, Xinyu Wang, Peng Lu, Xihao Yuan, Hanting Chen, Kai Han, Xinghao Chen, Chengjun Zhan, Hanlin Xu, Yichun Yin, Lifeng Shang, Feng Wen, Boxing Chen, Yufei Cui

The quadratic computational cost of traditional attention mechanisms poses a major bottleneck to the scalability and practical deployment of large language models (LLMs), particularly in long-context scenarios. To improve efficiency, existing approaches often enforce rigid structural constraints such as local attention windows. However, these strategies typically lead to substantial performance degradation on tasks requiring precise long-range recall. In this work, we propose MATCH, a scalable and efficient framework that augments sparsified attention mechanisms with dynamically integrated in-context information through an efficient retrieval system. Empirical results show that MATCH significantly improves the performance of sparse-attention models on both synthetic and real-world natural-language tasks. These findings highlight the versatility of MATCH as a general approach for enhancing in-context retrieval capabilities while maintaining the efficiency benefits of sparse attention architectures.

View free PDFSource page

Related papers

arxivcs.LGcs.AIcs.CL2026-07-20

A Controlled Study of Attention-Only Transformers

Henry Ndubuaku, Karen Mosoyan, Jakub Mroz, Noah Cylich, Satyajit Kumar, Parkirat Sandhu, et al.

Feed-forward networks hold two thirds of a transformer's non-embedding parameters, yet the architecture has not received a necessity test that controls parameters, compute, and depth at once. We pretrain attention-only decoder transformers (Simple Attention Networks, SANs) agains…

View free PDFSource page
arxivcs.CLcs.AIcs.LG2026-07-01

Logit-Contribution Scoring Identifies Non-Literal Retrieval Heads

Aryo Pradipta Gema, Beatrice Alex, Pasquale Minervini

In long-context use, large language models frequently synthesize answers from the meaning of a relevant context span rather than literally copy-pasting them. Identifying which attention heads perform this synthesis matters for interpreting long-context model behavior. Yet existin…

View free PDFSource page
arxivcs.LGcs.AIcs.CLcs.IR2026-07-11

Context by Distinct Information: An Auditable Dirichlet-Process Working Memory for Long, Redundant Context Streams

Siddharth Pal, Viktoria Rojkova

Context engineering decides what information a model carries forward, and current designs meter it in tokens: compressing the past into a bounded recurrent state, keeping a key-value entry for every token, or imposing a fixed budget through a window or eviction rule. All three ma…

View free PDFSource page
arxivcs.LGcs.AIcs.CL2026-07-02

Training Hybrid Block Diffusion Language Models with Partial Bidirectionality

Pranshu Chaturvedi, Parth Shroff, Tarun Suresh, Hangoo Kang, Kaiyue Wen

High-throughput long-context generation is one of the central challenges for large language models. Generation is typically memory-bandwidth-bound rather than compute-bound: each decoding step must stream the accumulated key/value (KV) cache from memory, so bandwidth demand grows…

View free PDFSource page
arxivcs.LGcs.AIcs.CL2026-07-15

Data-Efficient Adaptation of LLMs via Attention Head Reweighting

Tuomas Oikarinen, Zixiao Chen, Charlotte Siska, Tsui-Wei Weng, Chandan Singh, Jianfeng Gao

Learning effectively from limited data is critical in domains like security where labeled examples are scarce. Large language models (LLMs) have demonstrated some capabilities for data-efficient learning, especially through parameter-efficient adaptation methods, but continue to…

View free PDFSource page
arxivcs.CLcs.AIcs.LG2026-07-17

Induction in Both Directions: A Mechanistic Analysis of In-Context Learning in Masked Diffusion Language Models

Andy Catruna, Emilian Radoi

While the internal mechanisms of autoregressive (AR) transformers have been studied extensively, much less is known about diffusion language models (DLMs), an emerging alternative that generates text by iterative denoising. In this work, we study how DLMs implement induction, a m…

View free PDFSource page