# DAST: Context-Aware Compression in LLMs via Dynamic Allocation of Soft Tokens

arXiv: https://arxiv.org/abs/2502.11493 · HTML: https://mlsanity.com/paper/2502.11493/

## Key facts

- Cited by 22 papers in the map (data through 14 Oct 2026).
- Builds on 31 of its 51 references that are in the map.
- First posted on arXiv 17 Feb 2025.

## Builds on (most cited first, 30 of 31)

- [GPT-4 Technical Report](https://mlsanity.com/paper/2303.08774/) (2023)
- [LLaMA: Open and Efficient Foundation Language Models](https://mlsanity.com/paper/2302.13971/) (2023)
- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [Qwen Technical Report](https://mlsanity.com/paper/2309.16609/) (2023)
- [HuggingFace's Transformers: State-of-the-art Natural Language Processing](https://mlsanity.com/paper/1910.03771/) (2019)
- [LongBench: A Bilingual, Multitask Benchmark for Long Context Understanding](https://mlsanity.com/paper/2308.14508/) (2023)
- [MemGPT: Towards LLMs as Operating Systems](https://mlsanity.com/paper/2310.08560/) (2023)
- [SnapKV: LLM Knows What You are Looking for Before Generation](https://mlsanity.com/paper/2404.14469/) (2024)
- [LongLLMLingua: Accelerating and Enhancing LLMs in Long Context Scenarios via Prompt Compression](https://mlsanity.com/paper/2310.06839/) (2023)
- [Adapting Language Models to Compress Contexts](https://mlsanity.com/paper/2305.14788/) (2023)
- [LLMLingua-2: Data Distillation for Efficient and Faithful Task-Agnostic Prompt Compression](https://mlsanity.com/paper/2403.12968/) (2024)
- [BookSum: A Collection of Datasets for Long-form Narrative Summarization](https://mlsanity.com/paper/2105.08209/) (2021)
- [xRAG: Extreme Context Compression for Retrieval-augmented Generation with One Token](https://mlsanity.com/paper/2405.13792/) (2024)
- [Long Context Compression with Activation Beacon](https://mlsanity.com/paper/2401.03462/) (2024)
- [Natural Language Understanding and Inference with MLLM in Visual Question Answering: A Survey](https://mlsanity.com/paper/2411.17558/) (2024)
- [Benchmarking Multimodal Retrieval Augmented Generation with Dynamic VQA Dataset and Self-adaptive Planning Agent](https://mlsanity.com/paper/2411.02937/) (2024)
- [Are we ready for a new paradigm shift? A Survey on Visual Deep MLP](https://mlsanity.com/paper/2111.04060/) (2021)
- [The Past Mistake is the Future Wisdom: Error-driven Contrastive Probability Optimization for Chinese Spell Checking](https://mlsanity.com/paper/2203.00991/) (2022)
- [On the (In)Effectiveness of Large Language Models for Chinese Text Correction](https://mlsanity.com/paper/2307.09007/) (2023)
- [LatEval: An Interactive LLMs Evaluation Benchmark with Incomplete Information from Lateral Thinking Puzzles](https://mlsanity.com/paper/2308.10855/) (2023)
- [Global Attention Mechanism: Retain Information to Enhance Channel-Spatial Interactions](https://mlsanity.com/paper/2112.05561/) (2021)
- [Learning from the Dictionary: Heterogeneous Knowledge Guided Fine-tuning for Chinese Spell Checking](https://mlsanity.com/paper/2210.10320/) (2022)
- [Contrastive Learning with Hard Negative Entities for Entity Set Expansion](https://mlsanity.com/paper/2204.07789/) (2022)
- [Perception Compressor: A Training-Free Prompt Compression Framework in Long Context Scenarios](https://mlsanity.com/paper/2409.19272/) (2024)
- [Towards Real-World Writing Assistance: A Chinese Character Checking Benchmark with Faked and Misspelled Characters](https://mlsanity.com/paper/2311.11268/) (2023)
- [Let LLMs Take on the Latest Challenges! A Chinese Dynamic Question Answering Benchmark](https://mlsanity.com/paper/2402.19248/) (2024)
- [Cold-Start Recommendation towards the Era of Large Language Models (LLMs): A Comprehensive Survey and Roadmap](https://mlsanity.com/paper/2501.01945/) (2025)
- [SeqGPT: An Out-of-the-box Large Language Model for Open Domain Sequence Understanding](https://mlsanity.com/paper/2308.10529/) (2023)
- [A Frustratingly Easy Plug-and-Play Detection-and-Reasoning Module for Chinese Spelling Check](https://mlsanity.com/paper/2310.09119/) (2023)
- [One Example Shown, Many Concepts Known! Counterexample-Driven Conceptual Reasoning in Mathematical LLMs](https://mlsanity.com/paper/2502.10454/) (2025)

## Built on by (most cited first, 22 of 22)

- [Autoencoding-Free Context Compression for LLMs via Contextual Semantic Anchors](https://mlsanity.com/paper/2510.08907/) (2025)
- [COMI: Coarse-to-fine Context Compression via Marginal Information Gain](https://mlsanity.com/paper/2602.01719/) (2026)
- [Read As Human: Compressing Context via Parallelizable Close Reading and Skimming](https://mlsanity.com/paper/2602.01840/) (2026)
- [AdmTree: Compressing Lengthy Context with Adaptive Semantic Trees](https://mlsanity.com/paper/2512.04550/) (2025)
- [EvoConfig: Self-Evolving Multi-Agent Systems for Efficient Autonomous Environment Configuration](https://mlsanity.com/paper/2601.16489/) (2026)
- [Kwai Summary Attention Technical Report](https://mlsanity.com/paper/2604.24432/) (2026)
- [Gated Differentiable Working Memory for Long-Context Language Modeling](https://mlsanity.com/paper/2601.12906/) (2026)
- [CompLLM: Compression for Long Context Q&A](https://mlsanity.com/paper/2509.19228/) (2025)
- [SemantiCache: Efficient KV Cache Compression via Semantic Chunking and Clustered Merging](https://mlsanity.com/paper/2603.14303/) (2026)
- [From Token to Line: Enhancing Code Generation with a Long-Term Perspective](https://mlsanity.com/paper/2504.07433/) (2025)
- [Exposing and Defending the Achilles' Heel of Video Mixture-of-Experts](https://mlsanity.com/paper/2602.01369/) (2026)
- [PoC: Performance-oriented Context Compression for Large Language Models via Performance Prediction](https://mlsanity.com/paper/2603.19733/) (2026)
- [Beyond Position Bias: Shifting Context Compression from Position-Driven to Semantic-Driven](https://mlsanity.com/paper/2605.09463/) (2026)
- [Teaching According to Talents! Instruction Tuning LLMs with Competence-Aware Curriculum Learning](https://mlsanity.com/paper/2509.13790/) (2025)
- [Context Cartography: Toward Structured Governance of Contextual Space in Large Language Model Systems](https://mlsanity.com/paper/2603.20578/) (2026)
- [Mapping Text to Multiplex Graph: Prompt Compression as L'evy Walk-Guided Graph Pruning](https://mlsanity.com/paper/2607.01241/) (2026)
- [CL2GEC: A Multi-Discipline Benchmark for Continual Learning in Chinese Literature Grammatical Error Correction](https://mlsanity.com/paper/2509.13672/) (2025)
- [Experience-Driven Dynamic Exits for LLMs with Reinforcement Learning](https://mlsanity.com/paper/2606.03113/) (2026)
- [Dense Contexts Are Hard Contexts: Lexical Density Limits Effective Context in LLMs](https://mlsanity.com/paper/2606.06203/) (2026)
- [Planning with Transformers: Chain of Computation and Structured Context Windows](https://mlsanity.com/paper/2607.17710/) (2026)
- [Compression Beyond the Uncompressed: A Two-Stage Training Recipe for Soft Context Compression in RAG](https://mlsanity.com/paper/2609.05152/) (2026)
- [SuPER-Mamba: Surprisal Prioritization for Extended-Context Routing in Mamba](https://mlsanity.com/paper/2610.16636/) (2026)

## Abstract (arXiv)

Large Language Models (LLMs) face computational inefficiencies and redundant processing when handling long context inputs, prompting a focus on compression techniques. While existing semantic vector-based compression methods achieve promising performance, these methods fail to account for the intrinsic information density variations between context chunks, instead allocating soft tokens uniformly across context chunks. This uniform distribution inevitably diminishes allocation to information-critical regions. To address this, we propose Dynamic Allocation of Soft Tokens (DAST), a simple yet effective method that leverages the LLM's intrinsic understanding of contextual relevance to guide compression. DAST combines perplexity-based local information with attention-driven global information to dynamically allocate soft tokens to the informative-rich chunks, enabling effective, context-aware compression. Experimental results across multiple benchmarks demonstrate that DAST surpasses state-of-the-art methods.