paper-with-me

Papers

Long Context In-Context Compression by Getting to the Gist of Gisting

2025-04-11 · Aleksandar Petrov, Mark Sandler, Andrey Zhmoginov, Nolan Miller, Max Vladymyrov

Long context processing is critical for the adoption of LLMs, but existing methods often introduce architectural complexity that hinders their practical adoption. Gisting, an in-context compression method with no architectural modification to the decoder transformer, is a promising approach due to its simplicity and compatibility with existing frameworks. While effective for short instructions, we demonstrate that gisting struggles with longer contexts, with significant performance drops even at minimal compression rates. Surprisingly, a simple average pooling baseline consistently outperforms gisting. We analyze the limitations of gisting, including information flow interruptions, capacity limitations and the inability to restrict its attention to subsets of the context. Motivated by theoretical insights into the performance gap between gisting and average pooling, and supported by extensive experimentation, we propose GistPool, a new in-context compression method. GistPool preserves the simplicity of gisting, while significantly boosting its performance on long context compression tasks.

📄 PDF Abstract BibTeX arXiv:2504.08934

Code (0)

등록된 구현이 없습니다.

Tasks

Decoder

Methods 이 논문이 사용한 방법론

Softmax The Softmax output function transforms a previous layer's output into a vector of probabilities. It is commonly used for multiclass classification. Given an input vector $x$…
Attention 설명 없음
Average Pooling 설명 없음

Similar Papers 제목 키워드 기반

UniGist: Towards General and Hardware-aligned Sequence-level Long Context Compression

2025-09-19 · Chenlong Deng, Zhisong Zhang, Kelong Mao, Shuaiyi Li 외 arxiv

Large language models are increasingly capable of handling long-context inputs, but the memory overhead of key-value (KV) cache remains a major bottleneck for general-purpose deployment. While various compression strateg…

A Silver Bullet or a Compromise for Full Attention? A Comprehensive Study of Gist Token-based Context Compression

2024-12-23 · Chenlong Deng, Zhisong Zhang, Kelong Mao, Shuaiyi Li 외

In this work, we provide a thorough investigation of gist-based context compression methods to improve long-context processing in large language models. We focus on two key questions: (1) How well can these methods repla…

Retrieval-augmented Generation

Sentence-Anchored Gist Compression for Long-Context LLMs

2025-11-11 · Dmitrii Tarasov, Elizaveta Goncharova, Kuznetsov Andrey arxiv

This work investigates context compression for Large Language Models (LLMs) using learned compression tokens to reduce the memory and computational demands of processing long sequences. We demonstrate that pre-trained LL…

Compressing Lengthy Context With UltraGist

2024-05-26 · Peitian Zhang, Zheng Liu, Shitao Xiao, Ninglu Shao 외

Compressing lengthy context is a critical but technically challenging problem. In this paper, we propose a new method called UltraGist, which is distinguished for its high-quality compression of lengthy context due to th…

Few-Shot Learning

Gradual Forgetting: Logarithmic Compression for Extending Transformer Context Windows

2025-10-25 · Billy Dickson, Zoran Tiganj arxiv

Most approaches to long-context processing increase the complexity of the transformer's internal architecture by integrating mechanisms such as recurrence or auxiliary memory modules. In this work, we introduce an altern…