mlsanityData to 7 Oct 2026

PISCO: Pretty Simple Compression for Retrieval-Augmented Generation

arXiv 2501.16075 ↗First posted 27 Jan 2025cs.CL cs.AI cs.IRcited by 20

On the map: Retrieval-augmented generation → Generation and context integration

Abstract (arXiv)

Retrieval-Augmented Generation (RAG) pipelines enhance Large Language Models (LLMs) by retrieving relevant documents, but they face scalability issues due to high inference costs and limited context size. Document compression is a practical solution, but current soft compression methods suffer from accuracy losses and require extensive pretraining. In this paper, we introduce PISCO, a novel method that achieves a 16x compression rate with minimal accuracy loss (0-3%) across diverse RAG-based question-answering (QA) tasks. Unlike existing approaches, PISCO requires no pretraining or annotated data, relying solely on sequence-level knowledge distillation from document-based questions. With the ability to fine-tune a 7-10B LLM in 48 hours on a single A100 GPU, PISCO offers a highly efficient and scalable solution. We present comprehensive experiments showing that PISCO outperforms existing compression models by 8% in accuracy.

Key facts. Cited by 20 papers in the map (data through 7 Oct 2026). Placed in Generation and context integration, one of 15 directions in Retrieval-augmented generation. Builds on 30 of its 44 references that are in the map. First posted on arXiv 27 Jan 2025.

Builds on

30 of 44 references.

Closest ancestors · 5

Same direction, newest first

  1. 2024Inference Scaling for Long-Context Retrieval Augmented Generation6 Oct
  2. 2024Context Embeddings for Efficient Answer Generation in RAG12 Jul
  3. 2024xRAG: Extreme Context Compression for Retrieval-augmented Generation with One Token22 May
  4. 2023RECOMP: Improving Retrieval-Augmented LMs with Compression and Selective Augmentation6 Oct
  5. 2022When Not to Trust Language Models: Investigating Effectiveness of Parametric and Non-Parametric Memories20 Dec

Other references · 25

  1. 2021LoRA: Low-Rank Adaptation of Large Language Modelscited by 14,084
  2. 2024GPT-4o System Cardcited by 4,467
  3. 2018HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answeringcited by 3,686
  4. 2023Mistral 7Bcited by 3,065
  5. 2017TriviaQA: A Large Scale Distantly Supervised Challenge Dataset for Reading Comprehensioncited by 2,895
  6. 2021DeBERTaV3: Improving DeBERTa using ELECTRA-Style Pre-Training with Gradient-Disentangled Embedding Sharingcited by 1,219
  7. 2024Length-Controlled AlpacaEval: A Simple Way to Debias Automatic Evaluatorscited by 865
  8. 2016Sequence-Level Knowledge Distillationcited by 761
  9. 2020KILT: a Benchmark for Knowledge Intensive Language Taskscited by 571
  10. 2023Adapting Language Models to Compress Contextscited by 350
  11. 2024LLMLingua-2: Data Distillation for Efficient and Faithful Task-Agnostic Prompt Compressioncited by 343
  12. 2022Large Language Models Are Reasoning Teacherscited by 318
13 more
  1. 2022ASQA: Factoid Questions Meet Long-Form Answerscited by 231
  2. 2023In-context Autoencoder for Context Compression in a Large Language Modelcited by 190
  3. 2023Compressing Context to Enhance Inference Efficiency of Large Language Modelscited by 185
  4. 2024Long-Context LLMs Meet RAG: Overcoming Challenges for Long Inputs in RAGcited by 122
  5. 2022Prompt Compression and Contrastive Conditioning for Controllability and Toxicity Reduction in Language Modelscited by 96
  6. 2024Long-Context Language Modeling with Parallel Context Encodingcited by 84
  7. 2023Learning to Filter Context for Retrieval-Augmented Generationcited by 82
  8. 2018DuoRC: Towards Complex Language Understanding with Paraphrased Reading Comprehensioncited by 70
  9. 2024SPLADE-v3: New baselines for SPLADEcited by 69
  10. 2024500xCompressor: Generalized Prompt Compression for Large Language Modelscited by 45
  11. 2024BERGEN: A Benchmarking Library for Retrieval-Augmented Generationcited by 31
  12. 2024Stronger Models are NOT Stronger Teachers for Instruction Tuningcited by 16
  13. 2024Zero-shot cross-lingual transfer in instruction tuning of large language modelscited by 10

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 10; 2026: 10.

20 papers.

Sort:

By area:

  1. 2025Reducing Cost of LLM Agents with Trajectory Reductioncited by 44
  2. 2025SARA: Selective and Adaptive Retrieval-augmented Generation with Context Compressioncited by 9
  3. 2025CLaRa: Bridging Retrieval and Generation with Continuous Latent Reasoningcited by 9
  4. 2026Multi-Vector Index Compression in Any Modalitycited by 8
  5. 2025No Mean Feat: Simple, Strong Baselines for Context Compressioncited by 6
  6. 2025OSCAR: Online Soft Compression And Rerankingcited by 4
  7. 2025Adapting Large Language Models for Multi-Domain Retrieval-Augmented-Generationcited by 3
  8. 2025Privacy-Preserving Reasoning with Knowledge-Distilled Parametric Retrieval Augmented Generationcited by 3
  9. 2025ARC-Encoder: learning compressed text representations for large language modelscited by 3
  10. 2025Less Is More: Elevating RAG via Performance-Driven Context Compressioncited by 2
  11. 2025When More Retrieval Hurts: Retrieval-Augmented Code Review Generationcited by 2
  12. 2026XProvence: Zero-Cost Multilingual Context Pruning for Retrieval-Augmented Generationcited by 2
8 more
  1. 2026Cognitive Chunking for Soft Prompts: Accelerating Compressor Learning via Block-wise Causal Maskingcited by 1
  2. 2026Grounded Cache Routing for Retrieval-Augmented Generation: When Is It Safe to Reuse an Answer?cited by 1
  3. 2026Fixed RAG Compression Collapses Measured Reader Scalingcited by 1
  4. 2026A Unified Model and Document Representation for On-Device Retrieval-Augmented Generationcited by 0
  5. 2026RAGOCR: Optical Compression of Retrieval-Augmented Text via Visual Representationcited by 0
  6. 2026Compression Beyond the Uncompressed: A Two-Stage Training Recipe for Soft Context Compression in RAGcited by 0
  7. 2026REVA: Reusable Evidence View Aggregation for Context-Efficient RAG Servingcited by 0
  8. 2026BoundaryMORPH: Budgeted Reranking via Active Set Selection for Diffuse Retrievalcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2024Contextual Compression in Retrieval-Augmented Generation for Large Language Models: A Survey
  2. 2026Rethinking Soft Compression in Retrieval-Augmented Generation: A Query-Conditioned Selector Perspective
  3. 2024EXIT: Context-Aware Extractive Compression for Enhancing Retrieval-Augmented Generation
  4. 2025Knowledge Compression via Question Generation: Enhancing Multihop Document Retrieval without Fine-tuning
  5. 2025Question Decomposition for Retrieval-Augmented Generation
  6. 2024Lighter And Better: Towards Flexible Context Adaptation For Retrieval Augmented Generation
  7. 2024Long2RAG: Evaluating Long-Context & Long-Form Retrieval-Augmented Generation with Key Point Recall
  8. 2025Enhancing RAG Efficiency with Adaptive Context Compression