Fix the Structural Bottleneck: Context Compression via Explicit Information Transmission
cs.CL
Submitted: 2026-02-03
Updated: 2026-11-07
Code: https://github.com/Jiangnan0522/ComprExIT
License: http://creativecommons.org/licenses/by/4.0/
The gist: Long-context LLM agents often struggle with growing token, memory, and latency costs, making efficient context compression essential for practical deployment.
Terminology
Abstract
Long-context LLM agents often struggle with growing token, memory, and latency costs, making efficient context compression essential for practical deployment. Existing LLM-as-a-compressor methods remain noticeably inferior to using the full context. We find that this gap partly stems from their inability to preserve contextual information effectively. In this work, we revisit context compression from a structural perspective and identify two key bottlenecks in standard LLM-based compressors: limited coordination among compression tokens during information aggregation, and layerwise dilution that weakens useful signals from intermediate hidden states. To address these limitations, we propose ComprExIT, a new context compression framework based on explicit information transmission. ComprExIT adaptively selects features across frozen LLM layers, then allocates information from anchors to compression slots through a globally coordinated transport plan. Experiments on 12 datasets show that ComprExIT consistently outperforms strong soft-compression baselines, improving average F1 by up to 18.5%, while adding only ~1% trainable parameters and achieving more than 2x faster compression than the fastest baselines. The code will be released upon acceptance.
Sources
- EFPC: Towards Efficient and Flexible Prompt Compression
- The Llama 3 Herd of Models
- A Comprehensive Survey on Long Context Language Modeling
- SlimPajama-DC: Understanding Data Combinations for LLM Training
- Kimi K2.5: Visual Agentic Intelligence
- Qwen3.5-Omni Technical Report
- A Comprehensive Study of Knowledge Editing for Large Language Models
Related papers
- ALARM: Audio-Language Alignment for Reasoning Models
- ThinkSum: Probabilistic reasoning over sets using large language models
- Coherence boosting: When your pretrained language model is not paying enough attention
- From Facts to Insights: A Persona-Driven Dual Memory Framework and Dataset for Role-Playing Agents
- Improving General Role-Playing Agents via Psychology-Grounded Reasoning and Role-Aware Policy Optimization
- The Art of Asking: Multilingual Prompt Optimization for Synthetic Data