Context Compression

Follow this topic Ask about this topic

Context compression studies how to shrink long prompts, interaction histories, retrieved documents, or model memory without losing information needed for later tasks. Approaches include selecting or pruning text, rewriting it as summaries or structured records, encoding it into learned latent tokens, and adaptively retaining or retrieving history as an agent works. Methods are tested on long-context question answering, retrieval-augmented generation, coding and web agents, and long-horizon tool use, where accuracy, task success, latency, memory, and cost can all matter. The central tension is that aggressive compression can discard critical details or constraints, and fewer tokens do not always mean lower end-to-end cost or more reliable execution.

Latest papers

90 papers11 papers in the last four weeks, down 45% on the four weeks before. 0.1% of all new papers.

Diagnosing and Mitigating Long-Session Degradation in Full-Duplex Speech Models
Oct 26, 2026Junho Kim, CheHyun Lee, Donghun Kang +4Spoken Dialogue Systems Full-Duplex Spoken Dialogue Systems
IPAI, Seoul National University (SNU) · Department of ECE, SNU · Department of AI, University of Seoul +1
Bud ContextCompact 2.0: Context Compaction for Agent Success and Accuracy with Small Language Models and Constrained Context Windows
Oct 18, 2026Jithin VG, Ditto PSLLM Context Management LLM Context Compression
Bud Ecosystem
Maximising Transparency in Context Compaction Improves Retention of Facts and Commitments for Lightweight Coding Agents
Oct 14, 2026Felix WoolfordContext Compaction for LLM Agents Context Compression
Center for Human Nature, Artificial Intelligence, and Neuroscience (CHAIN), Hokkaido University
Compile the Table: Query-Calibrated Operator Compression for Tabular In-Context Learning
Oct 8, 2026Xu Zhao, Jiaming Zhao, Bin Zhao +1Context Compression Tabular ICL
REMORY: Learning Residual Memory for Context Compaction
Oct 8, 2026Hanchen Xia, Baoyou Chen, Yutang Ge +5LLM Agent Memory Memory-Augmented Language Model Agents
Shanghai Academy of AI for Science · Fudan University · Shanghai Jiao Tong University +5
ReFold: Training-Free Reversible Inter-Turn Context Folding for Long-Horizon Agents
Oct 6, 2026Yupeng Su, Jiayi Tian, Zheng Zhang +1LLM Inference Efficiency Long-Context Language Model Inference
University of California, Santa Barbara · Intel
FOCUS: Training-Free Decision-Preserving Context Compression for LLM Agents
Sep 29, 2026Shantanu Dixit, Anson Bastos, Xuchao Zhang +2Long-Horizon LLM Agents LLM Context Management
M365 Research, Microsoft
CADOC: Cache-Aware Dynamic Object Context for Long-Horizon Agents
Sep 29, 2026Junjie Yao, Zhangchen Zhou, Zhi-Qin John XuCost-Aware Inference Prefix Caching
School of Mathematical Sciences, Shanghai Jiao Tong University, Shanghai, China. · Institute of Natural Sciences, Shanghai Jiao Tong University, Shanghai, China.
Adapting Context Compression for Long-Horizon Agents with Counterfactual Continuations
Sep 29, 2026Guanghui Min, Liang Wu, Mingjia Shi +4Counterfactual Evaluation LLM Agent Evaluation
University of Virginia · Nokia
KV-streams for Efficient Compaction in Agentic Reinforcement Learning
Sep 28, 2026Emiliano Penaloza, Dane Malenfant, Dheeraj Vattikonda +15Recurrent Language Models Agentic RL
Mila · Microsoft · Université de Montréal +9
Scalable In-Context Reinforcement Learning with Recurrent Algorithm Distillation
Sep 28, 2026Yuanqing Ma, Zhenrui Zheng, Chenjun XiaoTransformer In-Context RL
The Chinese University of Hong Kong, Shenzhen
Instance-Adaptive Prompts as Context for Time-Series Foundation Models
Sep 28, 2026Zehao Xiao, Shifeng Xie, Lei Zan +5Time Series Forecasting Prompt Tuning
Huawei Noah’s Ark Lab, Paris, France · LIPADE, Universit´e Paris Cit´e, Paris, France · Huawei Noah’s Ark Lab, Shenzhen, China
Downstream-Aware Context Selection for Online In-Context Reinforcement Learning
Sep 27, 2026Ruihan A. Li, Shangtong Zhang, Rohan ChandraIn-Context RL In-Context Learning
University of Illinois Urbana-Champaign · University of Virginia
Beyond Token Savings: A Systematic Study of Context Compression in LLM Agents
Sep 26, 2026Ritul Satish, Prasoon Sinha, Akiho Kawada +1Context Compaction for LLM Agents Context Compression
The University of Texas at Austin
Compress What You See, Not What You Say: Anchored Context Distillation for Latent-Observation Software Engineering Agents
Sep 25, 2026Zhensheng Zou, Guoqing Wang, Dan HaoLLM Context Management Context Compression
Peking University
Evaluating Budgeted Context Projection with Unexecuted Companion Runs
Sep 25, 2026Guangzhe ZhangLLM Context Management LLM Agent Evaluation
Independent AI Researcher
When Can Agents Forget Their Reasoning? ICLR for Long-Horizon Agent Context Compression
Sep 24, 2026Mingxuan Wang, Fei Luo, Bo Wang +6Long-Context Modeling Long-Horizon LLM Agents
TierFlow Team · Gaoling School of Artificial Intelligence, Renmin University of China · Tsinghua University
MORSE: Multi-Context Ordering via Reverse Scoring for Evidence-Preserving Compression
Sep 23, 2026Ke Wan, Yifan Wang, Liheng Lai +1Retrieval-Augmented Generation Evidence Retrieval
Department of Computer Science University of Virginia Charlottesville, VA, USA · AfterQuery
Stable Geometry with Divergent Task Evidence for Efficient Long-Horizon Agent Compression
Sep 23, 2026Mingxuan Wang, Fei Luo, Bo Wang +6Agent Memory Management Context Compression
TierFlow Team · Gaoling School of Artificial Intelligence, Renmin University of China · Tsinghua University
StateComp: Learning When to Compress History in Long Horizon Agents
Sep 23, 2026Mingxuan Wang, Hongyue Chen, Yinglong Guo +6Context Compression Long-Horizon LLM Agents
TierFlow Team · Gaoling School of Artificial Intelligence, Renmin University of China · Tsinghua University
Memory Control Signals Emerge Before Action in Long Horizon Agents
Sep 23, 2026Mingxuan Wang, Guorun Yao, Fei Luo +6Agent Memory Management Long-Horizon LLM Agents
TierFlow Team · Gaoling School of Artificial Intelligence, Renmin University of China · Tsinghua University
DRSR: Learning Set-Level Deletion Risk for Efficient Long-Horizon Agents
Sep 23, 2026Mingxuan Wang, Bo Wang, Fei Luo +6Context Compression Long-Horizon LLM Agents
TierFlow Team · Gaoling School of Artificial Intelligence, Renmin University of China · Tsinghua University
CliffCompaction: Cost-Efficient Compaction for Long-Horizon Coding Agents
Sep 22, 2026Trang Nguyen, Eulrang Cho, Bingqing Chen +1Continual Learning for LLM Agents Test-Time Scaling
Carnegie Mellon University · Bosch Center for AI
Compressing Long Context into Answer-Aligned Memory Embeddings for LLM Inference
Sep 22, 2026Md Mostafizer Rahman, Md Faizul Ibne Amin, Md Shahajada Mia +2LLM Inference Efficiency Long-Context Language Model Inference
Lucy Family Institute for Data & Society, University of Notre Dame, IN, USA · The University of Aizu, Aizuwakamatsu, Japan · Department of Applied and Computational Mathematics and Statistics, University of Notre Dame, IN, USA
Correct Now, Insufficient Later: Auditing Update Sufficiency in Context Compression
Sep 17, 2026Guangzhe ZhangAgent Memory Counterfactual Evaluation
Independent AI Researcher
Protocol-Preserving Context Trimming for Agentic Workflows: Benefits, Failure Regimes, and Budget Guardrails
Sep 15, 2026Harish GaggarLLM Context Management LLM Agent Workflow Optimization
Intuit Credit Karma, Oakland, California, USA
FlexComp: One Model for Every Ratio in Context Compression
Sep 10, 2026Kaiyan Zhao, Zhongtao Miao, Akiko Aizawa +1LLM Inference Acceleration Context Compression
The University of Tokyo · National Institute of Informatics
AttnCompress: Dynamic Attention-Guided Trajectory Compression for Software Engineering Agents
Sep 8, 2026Zhengran Zeng, Yixin Li, Rui Xie +2Software Engineering Agents Automated Software Engineering
Peking University, China
RAGMark: A Comprehensive Framework for Benchmarking Retrieval-Augmented Generation Systems
Sep 4, 2026Zlatan Feric, Amir Taherin, Bin Ren +3Retrieval-Augmented Generation RAG Evaluation
Northeastern University, Boston, MA, USA · College of William & Mary, VA, USA
LatentPress: Context Compression Beyond Text and Vision
Sep 1, 2026Yuanda Xu, Zhengze Zhou, Hejian Sang +3LLM Inference Acceleration LLM Context Compression
Iowa State University · Cornell University

Oct 26, 2026J/K move · Enter open · S save

Diagnosing and Mitigating Long-Session Degradation in Full-Duplex Speech Models

Junho Kim, CheHyun Lee, Donghun Kang, Youngho Cho+3

IPAI, Seoul National University (SNU) · Department of ECE, SNU · Department of AI, University of Seoul+1

Spoken Dialogue Systems Full-Duplex Spoken Dialogue Systems Context Compression

Open paper AskPDF arXiv

End-to-end full-duplex spoken dialogue models enable natural, real-time spoken interactions with interruptions, backchannels, and overlapping speech. Prior evaluations have assessed diverse interaction capabilities and identified some degradation patterns across conversational turns. However, how model behavior evolves over extended sessions remains insufficiently understood. In this work, we address this gap by systematically characterizing long-session degradation and developing efficient methods to mitigate it. We introduce a closed-loop evaluation framework for long-session stability and apply it to several publicly available full-duplex speech models, revealing degradation over time in responsiveness, spoken content quality, and voice consistency. To mitigate this degradation, we propose a self-distillation-based training framework for lightweight LoRA adaptation that improves long-session stability with a small amount of additional training. We further introduce a two-resolution context (TRC) architecture that preserves recent interactions in full detail while compressing older interaction history into key-value representations. We evaluate our methods on three models with distinct architectures, assessing long-session stability, information retention, and computational efficiency. Across all three models, we show that LoRA adaptation mitigates performance degradation over ten-minute sessions, while adding TRC extends these gains beyond the training session length and reduces KV-cache growth by more than 90%.