Evaluating Zero-Shot Long-Context LLM Compression

• URL: http://arxiv.org/abs/2406.06773v2
• Date: Thu, 13 Feb 2025 17:50:39 GMT
• Title: Evaluating Zero-Shot Long-Context LLM Compression
• Authors: Chenyu Wang, Yihan Wang, Kai Li,
• Abstract summary: This report examines the effectiveness of zero-shot compression techniques on large language models (LLMs) under long-context.<n>Due to limited computational resources, our experiments were conducted only on LLaMA-2-7B-32K.
• Score: 22.14193187475446
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: This study evaluates the effectiveness of zero-shot compression techniques on large language models (LLMs) under long-context. We identify the tendency for computational errors to increase under long-context when employing certain compression methods. We propose a hypothesis to explain the varied behavior of different LLM compression techniques and explore remedies to mitigate the performance decline observed in some techniques under long-context. This is a course report for COS 598D Machine Learning and Systems by Prof. Kai Li at Princeton University. Due to limited computational resources, our experiments were conducted only on LLaMA-2-7B-32K.

Related papers

• Gradually Compacting Large Language Models for Reasoning Like a Boiling Frog [72.4168434368873]
  Large Language Models (LLMs) have demonstrated impressive reasoning capabilities, but their substantial size often demands significant computational resources.<n>We propose a gradual compacting method that divides the compression process into multiple fine-grained iterations.<n>This iterative approach-reminiscent of the "boiling frog" effect-enables the model to be progressively compressed without abrupt performance loss.
  arXiv  Detail & Related papers  (2026-02-04T06:56:52Z)
• Sentence-Anchored Gist Compression for Long-Context LLMs [3.4406991639518307]
  We show that pre-trained Large Language Models can be fine-tuned to compress their context by factors of 2x to 8x without significant performance degradation.<n>Our method achieves results on par with alternative compression techniques while attaining higher compression ratios.
  arXiv  Detail & Related papers  (2025-11-11T11:34:32Z)
• Investigating Structural Pruning and Recovery Techniques for Compressing Multimodal Large Language Models: An Empirical Study [64.26593350748401]
  Multimodal Large Language Models (MLLMs) demonstrate impressive capabilities.<n>Current parameter reduction techniques primarily involve training MLLMs from Small Language Models (SLMs)<n>We propose to directly compress existing MLLMs through structural pruning combined with efficient recovery training.
  arXiv  Detail & Related papers  (2025-07-28T11:57:52Z)
• Compression Laws for Large Language Models [20.62274005080048]
  We introduce compression laws for language language models (LLMs) We empirically examine the effects of structured model compression on LLMs through over $1000$ experiments. Our findings indicate that the test cross-entropy loss increases quadratically with the compression ratio.
  arXiv  Detail & Related papers  (2025-04-06T03:39:34Z)
• Efficient Long Context Language Model Retrieval with Compression [57.09163579304332]
  Long Context Language Models (LCLMs) have emerged as a new paradigm to perform Information Retrieval (IR) We propose a new compression approach tailored for LCLM retrieval, which is trained to maximize the retrieval performance while minimizing the length of the compressed passages. We show that CoLoR improves the retrieval performance by 6% while compressing the in-context size by a factor of 1.91.
  arXiv  Detail & Related papers  (2024-12-24T07:30:55Z)
• Selection-p: Self-Supervised Task-Agnostic Prompt Compression for Faithfulness and Transferability [67.77534983324229]
  In this paper, we investigate the ability of Large Language Models to develop a unified compression method that discretizes uninformative tokens. Experiments show Selection-p achieves state-of-the-art performance across numerous classification tasks. It exhibits superior transferability to different models compared to prior work.
  arXiv  Detail & Related papers  (2024-10-15T17:05:25Z)
• LanguaShrink: Reducing Token Overhead with Psycholinguistics [8.123272461141815]
  LanguaShrink is a prompt compression framework for large language models. It reduces prompt length while preserving essential information. Compared to existing prompt compression methods, LanguaShrink improves end-to-end latency by 1.43 times.
  arXiv  Detail & Related papers  (2024-09-01T22:09:20Z)
• In-Context Former: Lightning-fast Compressing Context for Large Language Model [48.831304302467004]
  In this paper, we propose a new approach to compress the long input contexts of Transformer-based large language models (LLMs) We use the cross-attention mechanism and a small number of learnable digest tokens to condense information from the contextual word embeddings. Experimental results indicate that our method requires only 1/32 of the floating-point operations of the baseline during compression and improves processing speed by 68 to 112 times.
  arXiv  Detail & Related papers  (2024-06-19T15:14:55Z)
• Feature-based Low-Rank Compression of Large Language Models via Bayesian Optimization [40.15915011575071]
  Low-rank compression is a promising technique to reduce non-essential parameters in large language models. We conduct empirical research on the low-rank characteristics of large models. We propose a low-rank compression method suitable for large language models.
  arXiv  Detail & Related papers  (2024-05-17T08:27:12Z)
• What Happens When Small Is Made Smaller? Exploring the Impact of Compression on Small Data Pretrained Language Models [2.2871867623460216]
  This paper investigates the effectiveness of pruning, knowledge distillation, and quantization on an exclusively low-resourced, small-data language model, AfriBERTa. Through a battery of experiments, we assess the effects of compression on performance across several metrics beyond accuracy.
  arXiv  Detail & Related papers  (2024-04-06T23:52:53Z)
• Compressing LLMs: The Truth is Rarely Pure and Never Simple [90.05366363633568]
  Knowledge-Intensive Compressed LLM BenchmarK aims to redefine the evaluation protocol for compressed Large Language Models. LLM-KICK unveils many favorable merits and unfortunate plights of current SoTA compression methods. LLM-KICK is designed to holistically access compressed LLMs' ability for language understanding, reasoning, generation, in-context retrieval, in-context summarization, etc.
  arXiv  Detail & Related papers  (2023-10-02T17:42:37Z)
• Do Compressed LLMs Forget Knowledge? An Experimental Study with Practical Implications [63.29358103217275]
  Large Language Models (LLMs) often leads to reduced performance, especially for knowledge-intensive tasks. We propose two conjectures on the nature of the damage: one is certain knowledge being forgotten (or erased) after compression. We introduce a variant called Inference-time Dynamic Prompting (IDP) that can effectively increase prompt diversity without incurring any inference overhead.
  arXiv  Detail & Related papers  (2023-10-02T03:12:06Z)
• Approximating Human-Like Few-shot Learning with GPT-based Compression [55.699707962017975]
  We seek to equip generative pre-trained models with human-like learning capabilities that enable data compression during inference. We present a novel approach that utilizes the Generative Pre-trained Transformer (GPT) to approximate Kolmogorov complexity.
  arXiv  Detail & Related papers  (2023-08-14T05:22:33Z)
• Compressing Large Sample Data for Discriminant Analysis [78.12073412066698]
  We consider the computational issues due to large sample size within the discriminant analysis framework. We propose a new compression approach for reducing the number of training samples for linear and quadratic discriminant analysis.
  arXiv  Detail & Related papers  (2020-05-08T05:09:08Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.