SepSeq: A Training-Free Framework for Long Numerical Sequence Processing in LLMs (2026.findings-acl)
Copied to clipboard
Jie Sun, Yu Liu, Lu Han, Qiwen Deng, Xiang Shu, Yang Xiao, Lintao Ma, Xingyu Lu, Jun Zhou, Pengfei Liu, Jiancan Wu, Xiang Wang
| Challenge: | Existing large-scale large-context models suffer from performance degradation when processing long numerical sequences. |
| Approach: | They propose a framework to mitigate attention dispersion by strategically inserting separator tokens into the model to recalibrat attention to local segments while preserving global context. |
| Outcome: | The proposed framework improves accuracy and reduces inference token consumption by 16.4% on 9 widely-adopted LLMs. |
Similar Papers
LLM×MapReduce: Simplified Long-Sequence Processing using Large Language Models (2025.acl-long)
Copied to clipboard
Zihan Zhou, Chong Li, Xinyi Chen, Shuo Wang, Yu Chao, Zhili Li, Haoyu Wang, Qi Shi, Zhixing Tan, Xu Han, Xiaodong Shi, Zhiyuan Liu, Maosong Sun
| Challenge: | Existing studies have focused on extending the context length of large language models (LLMs) due to their quadratic computational complexity and a lack of high-quality long training examples, most LLMs are trained with a limited window size. |
| Approach: | They propose a training-free framework that enables large language models to effectively process long texts using a divide-and-conquer strategy for comprehensive document understanding. |
| Outcome: | The proposed framework outperforms open-source and commercial long-context LLMs and is compatible with several models. |
FocusLLM: Precise Understanding of Long Context by Dynamic Condensing (2025.acl-long)
Copied to clipboard
Zhenyu Li, Yike Zhang, Tengyu Pan, Yutao Sun, Zhichao Duan, Junjie Fang, Rong Han, Zixuan Wang, Jianyong Wang
| Challenge: | Existing context condensing methods cannot accurately understand the full context, as there is a considerable amount of information loss in the condensed process. |
| Approach: | They propose a framework to extend the fixed context length of any decoder-only LLM by distilling crucial information from long sequences. |
| Outcome: | The proposed framework extends the fixed context length of any decoder-only LLM, allowing it to focus on relevant information from very long sequences. |
Focus-dLLM: Accelerating Long-Context Diffusion LLM Inference via Confidence-Guided Context Focusing (2026.acl-long)
Copied to clipboard
| Challenge: | Existing methods for estimating attention importance for tokens are ineffective . dLLMs require bidirectional attention, which limits inference efficiency . |
| Approach: | They propose a training-free attention sparsification framework for efficient long-context inference . they propose 'sink-aware pruning strategy' to accurately estimate and remove redundant computation . |
| Outcome: | The proposed approach offers 29 lossless speedup under 32K context length. |
LongRecipe: Recipe for Efficient Long Context Generalization in Large Language Models (2025.acl-long)
Copied to clipboard
Zhiyuan Hu, Yuliang Liu, Jinman Zhao, Suyuchen Wang, WangYan WangYan, Wei Shen, Qing Gu, Anh Tuan Luu, See-Kiong Ng, Zhiwei Jiang, Bryan Hooi
| Challenge: | Large language models face significant challenges in handling long-context tasks because of their limited effective context window size during pretraining, which restricts their ability to generalize over extended sequences. |
| Approach: | They propose a training strategy for extending the context window of LLMs including impactful token analysis, position index transformation, and training optimization strategies. |
| Outcome: | Experiments on three types of LLMs show that LongRecipe can utilize long sequences while requiring only 30% of the target context window size. |
LongAttn: Selecting Long-context Training Data via Token-level Attention (2025.findings-acl)
Copied to clipboard
| Challenge: | Existing methods to select long-context data often rely on sentence-level analysis, which can be greatly optimized in both performance and efficiency. |
| Approach: | They propose a token-level framework which quantifies long-range dependencies for LLMs by calculating token-based dependency strength and distribution uniformity of token scores. |
| Outcome: | The proposed framework quantifies long-range dependencies, enabling more accurate and efficient data selection. |
LCIRC: A Recurrent Compression Approach for Efficient Long-form Context and Query Dependent Modeling in LLMs (2025.naacl-long)
Copied to clipboard
| Challenge: | Large language models excel in generating coherent and contextually rich outputs, but their capacity to handle long-form contexts is limited by fixed-length position embeddings. |
| Approach: | They propose a method that enables the efficient processing long-form sequences beyond the model’s length limit through recurrent compression without retraining the entire model. |
| Outcome: | The proposed method significantly improves LLM’s ability to manage extended contexts, making it well-suited for tasks that require both comprehensive context understanding and query relevance. |
EpMAN: Episodic Memory AttentioN for Generalizing to Longer Contexts (2025.acl-long)
Copied to clipboard
Subhajit Chaudhury, Payel Das, Sarathkrishna Swaminathan, Georgios Kollias, Elliot Nelson, Khushbu Pahwa, Tejaswini Pedapati, Igor Melnyk, Matthew Riemer
| Challenge: | Recent advances in Large Language Models (LLMs) have yielded impressive successes on many language tasks, but efficient processing of long contexts remains a significant challenge. |
| Approach: | They propose a method for processing long contexts in an episodic memory module while holistically attending to semantically-relevant context chunks. |
| Outcome: | The proposed method outperforms baseline decoders on multiple long-context recall and question-answering benchmarks on 16k to 256k tokens. |
DELTA: Dynamic Layer-Aware Token Attention for Efficient Long-Context Reasoning (2026.findings-acl)
Copied to clipboard
| Challenge: | Large reasoning models generate long chains of intermediate steps, but their inference cost is dominated by decoding, where each new token must attend to the entire growing sequence. |
| Approach: | They propose a training-free sparse attention mechanism that reduces inference cost by evicting entries from the key-value cache. |
| Outcome: | The proposed model matches or surpasses full attention on reasoning benchmarks . it reduces the number of attended tokens by up to 4.25 and delivers 1.54 speedup . |
LongSpec: Long-Context Lossless Speculative Decoding with Efficient Drafting and Verification (2026.acl-long)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) can process extremely long contexts, requiring efficient inference over extended inputs. |
| Approach: | They propose a model that uses a constant-sized key-value cache to train long-context models. |
| Outcome: | Experimental results show that LongSpec achieves 3.26x speedup over strong Flash Attention baselines and 2.34x wall clock time on four math reasoning tasks. |
FoldMoE: Efficient Long Sequence MoE Training via Attention-MoE Pipelining (2025.acl-long)
Copied to clipboard
| Challenge: | Existing approaches to training LLMs with Mixture-of-Experts (MoE) architecture on long sequences are limited by the insufficient computation. |
| Approach: | They propose a MoE training system that enables token-level overlapping across entire Transformer blocks through novel attention-MoE pipelining. |
| Outcome: | The proposed system achieves 1.49x and 2.72x speedup over state-of-the-art token-level overlapping and non-overlapping baselines on GPT-MoE models with sequences up to 32K tokens. |