Cover Image

Fix the Structural Bottleneck: Context Compression via Explicit Information Transmission

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Ye, Jiangnan, Yan, Hanqi, Shen, Zhenyi, Chang, Heng, Mao, Ye, He, Yulan
Format:	Preprint
Published:	2026
Subjects:	Computation and Language
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

CODI: Compressing Chain-of-Thought into Continuous Space via Self-Distillation
by: Shen, Zhenyi, et al.
Published: (2025)

Explainable Recommender with Geometric Information Bottleneck
by: Yan, Hanqi, et al.
Published: (2023)

Addressing Order Sensitivity of In-Context Demonstration Examples in Causal Language Models
by: Xiang, Yanzheng, et al.
Published: (2024)

Beyond Prompting: An Efficient Embedding Framework for Open-Domain Question Answering
by: Hu, Zhanghao, et al.
Published: (2025)

The Mystery of In-Context Learning: A Comprehensive Survey on Interpretation and Analysis
by: Zhou, Yuxiang, et al.
Published: (2023)

WhenLoss: Diagnosing Write and Retrieval Bottlenecks in Long-Context Memory Systems
by: Yu, Jiangnan, et al.
Published: (2026)

Beyond Perplexity: Let the Reader Select Retrieval Summaries via Spectrum Projection Score
by: Hu, Zhanghao, et al.
Published: (2025)

Weak Reward Model Transforms Generative Models into Robust Causal Event Extraction Systems
by: da Silva, Italo Luis, et al.
Published: (2024)

When Thinking Backfires: Mechanistic Insights Into Reasoning-Induced Misalignment
by: Yan, Hanqi, et al.
Published: (2025)

Mirror: A Multiple-perspective Self-Reflection Method for Knowledge-rich Reasoning
by: Yan, Hanqi, et al.
Published: (2024)

GraphMind: Interactive Novelty Assessment System for Accelerating Scientific Discovery
by: da Silva, Italo Luis, et al.
Published: (2025)

Soft Reasoning: Navigating Solution Spaces in Large Language Models through Controlled Embedding Exploration
by: Zhu, Qinglin, et al.
Published: (2025)

RL from Teacher-Model Refinement: Gradual Imitation Learning for Machine Translation
by: Lee, Dongyub Jude, et al.
Published: (2025)

QUITO-X: A New Perspective on Context Compression from the Information Bottleneck Theory
by: Wang, Yihang, et al.
Published: (2024)

SSA: Sparse Sparse Attention by Aligning Full and Sparse Attention Outputs in Feature Space
by: Shen, Zhenyi, et al.
Published: (2025)

Encourage or Inhibit Monosemanticity? Revisit Monosemanticity from a Feature Decorrelation Perspective
by: Yan, Hanqi, et al.
Published: (2024)

BOW: Reinforcement Learning for Bottlenecked Next Word Prediction
by: Shen, Ming, et al.
Published: (2025)

SciReplicate-Bench: Benchmarking LLMs in Agent-driven Algorithmic Reproduction from Research Papers
by: Xiang, Yanzheng, et al.
Published: (2025)

Compressing Lengthy Context With UltraGist
by: Zhang, Peitian, et al.
Published: (2024)

Beyond RAG for Agent Memory: Retrieval by Decoupling and Aggregation
by: Hu, Zhanghao, et al.
Published: (2026)

Counterfactual Generation with Identifiability Guarantees
by: Yan, Hanqi, et al.
Published: (2024)

Stop the Flip-Flop: Context-Preserving Verification for Fast Revocable Diffusion Decoding
by: Xiang, Yanzheng, et al.
Published: (2026)

The Reasoning Bottleneck in Graph-RAG: Structured Prompting and Context Compression for Multi-Hop QA
by: Zarrinkia, Yasaman, et al.
Published: (2026)

Long Context Compression with Activation Beacon
by: Zhang, Peitian, et al.
Published: (2024)

GMSA: Enhancing Context Compression via Group Merging and Layer Semantic Alignment
by: Tang, Jiwei, et al.
Published: (2025)

StreamingDialogue: Prolonged Dialogue Learning via Long Context Compression with Minimal Losses
by: Li, Jia-Nan, et al.
Published: (2024)

Fine-Grained Modeling of Narrative Context: A Coherence Perspective via Retrospective Questions
by: Xu, Liyan, et al.
Published: (2024)

COMI: Coarse-to-fine Context Compression via Marginal Information Gain
by: Tang, Jiwei, et al.
Published: (2026)

Extracting Event Temporal Relations via Hyperbolic Geometry
by: Tan, Xingwei, et al.
Published: (2021)

Learning Intrinsic Dimension via Information Bottleneck for Explainable Aspect-based Sentiment Analysis
by: Cheng, Zhenxiao, et al.
Published: (2024)

Sentinel: Decoding Context Utilization via Attention Probing for Efficient LLM Context Compression
by: Zhang, Yong, et al.
Published: (2025)

Compressing KV Cache for Long-Context LLM Inference with Inter-Layer Attention Similarity
by: Ma, Da, et al.
Published: (2024)

Uncertainty Estimation on Sequential Labeling via Uncertainty Transmission
by: He, Jianfeng, et al.
Published: (2023)

SCOPE: Optimizing Key-Value Cache Compression in Long-context Generation
by: Wu, Jialong, et al.
Published: (2024)

CCF: A Context Compression Framework for Efficient Long-Sequence Language Modeling
by: Li, Wenhao, et al.
Published: (2025)

VQ-Logits: Compressing the Output Bottleneck of Large Language Models via Vector Quantized Logits
by: Shao, Jintian, et al.
Published: (2025)

From Context to EDUs: Faithful and Structured Context Compression via Elementary Discourse Unit Decomposition
by: Zhou, Yiqing, et al.
Published: (2025)

Unveiling In-Context Learning: A Coordinate System to Understand Its Working Mechanism
by: Zhao, Anhao, et al.
Published: (2024)

Linguistic Structure from a Bottleneck on Sequential Information Processing
by: Futrell, Richard, et al.
Published: (2024)

In-Context Former: Lightning-fast Compressing Context for Large Language Model
by: Wang, Xiangfeng, et al.
Published: (2024)