Unveiling and Harnessing Hidden Attention Sinks: Enhancing Large Language Models without Training through Attention Calibration

• URL: http://arxiv.org/abs/2406.15765v1
• Date: Sat, 22 Jun 2024 07:00:43 GMT
• Title: Unveiling and Harnessing Hidden Attention Sinks: Enhancing Large Language Models without Training through Attention Calibration
• Authors: Zhongzhi Yu, Zheng Wang, Yonggan Fu, Huihong Shi, Khalid Shaikh, Yingyan Celine Lin,
• Abstract summary: We aim to provide a more profound understanding of the existence of attention sinks within large language models (LLMs) We propose a training-free Attention Technique (ACT) that automatically optimize the attention distributions on the fly during inference in an input-adaptive manner. ACT achieves an average improvement of up to 7.30% in accuracy across different datasets when applied to Llama-30B.
• Score: 15.36841874118801
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Attention is a fundamental component behind the remarkable achievements of large language models (LLMs). However, our current understanding of the attention mechanism, especially regarding how attention distributions are established, remains limited. Inspired by recent studies that explore the presence of attention sink in the initial token, which receives disproportionately large attention scores despite their lack of semantic importance, this work delves deeper into this phenomenon. We aim to provide a more profound understanding of the existence of attention sinks within LLMs and to uncover ways to enhance the achievable accuracy of LLMs by directly optimizing the attention distributions, without the need for weight finetuning. Specifically, this work begins with comprehensive visualizations of the attention distributions in LLMs during inference across various inputs and tasks. Based on these visualizations, to the best of our knowledge, we are the first to discover that (1) attention sinks occur not only at the start of sequences but also within later tokens of the input, and (2) not all attention sinks have a positive impact on the achievable accuracy of LLMs. Building upon our findings, we propose a training-free Attention Calibration Technique (ACT) that automatically optimizes the attention distributions on the fly during inference in an input-adaptive manner. Extensive experiments validate that ACT consistently enhances the accuracy of various LLMs across different applications. Specifically, ACT achieves an average improvement of up to 7.30% in accuracy across different datasets when applied to Llama-30B. Our code is available at https://github.com/GATECH-EIC/ACT.

Related papers

• Learn from Downstream and Be Yourself in Multimodal Large Language Model Fine-Tuning [104.27224674122313]
  Fine-tuning MLLM has become a common practice to improve performance on specific downstream tasks. To balance the trade-off between generalization and specialization, we propose measuring the parameter importance for both pre-trained and fine-tuning distributions.
  arXiv  Detail & Related papers  (2024-11-17T01:16:37Z)
• Attention Tracker: Detecting Prompt Injection Attacks in LLMs [62.247841717696765]
  Large Language Models (LLMs) have revolutionized various domains but remain vulnerable to prompt injection attacks. We introduce the concept of the distraction effect, where specific attention heads shift focus from the original instruction to the injected instruction. We propose Attention Tracker, a training-free detection method that tracks attention patterns on instruction to detect prompt injection attacks.
  arXiv  Detail & Related papers  (2024-11-01T04:05:59Z)
• When Attention Sink Emerges in Language Models: An Empirical View [39.36282162213973]
  Language Models (LMs) assign significant attention to the first token, even if it is not semantically important. This phenomenon has been widely adopted in applications such as streaming/long context generation, KV cache optimization, inference acceleration, model quantization, and others. We first demonstrate that attention sinks exist universally in LMs with various inputs, even in small models.
  arXiv  Detail & Related papers  (2024-10-14T17:50:28Z)
• Found in the Middle: Calibrating Positional Attention Bias Improves Long Context Utilization [97.84156490765457]
  Large language models (LLMs) struggle to capture relevant information located in the middle of their input. This phenomenon has been known as the lost-in-the-middle problem. We show found-in-the-middle achieves better performance in locating relevant information within a long context.
  arXiv  Detail & Related papers  (2024-06-23T04:35:42Z)
• Extending Token Computation for LLM Reasoning [5.801044612920816]
  Large Language Models (LLMs) are pivotal in advancing natural language processing. LLMs often struggle with complex reasoning tasks due to inefficient attention distributions. We introduce a novel method for extending computed tokens in the Chain-of-Thought process, utilizing attention mechanism optimization.
  arXiv  Detail & Related papers  (2024-03-22T03:23:58Z)
• Fortify the Shortest Stave in Attention: Enhancing Context Awareness of Large Language Models for Effective Tool Use [74.72150542395487]
  An inherent waveform pattern in the attention allocation of large language models (LLMs) significantly affects their performance in tasks demanding a high degree of context awareness. To address this issue, we propose a novel inference method named Attention Buckets.
  arXiv  Detail & Related papers  (2023-12-07T17:24:51Z)
• Paying More Attention to Self-attention: Improving Pre-trained Language Models via Attention Guiding [35.958164594419515]
  Pre-trained language models (PLM) have demonstrated their effectiveness for a broad range of information retrieval and natural language processing tasks. As the core part of PLM, multi-head self-attention is appealing for its ability to jointly attend to information from different positions. We propose two kinds of attention guiding methods, i.e., map discrimination guiding (MDG) and attention pattern decorrelation guiding (PDG)
  arXiv  Detail & Related papers  (2022-04-06T16:22:02Z)
• Alignment Attention by Matching Key and Query Distributions [48.93793773929006]
  This paper introduces alignment attention that explicitly encourages self-attention to match the distributions of the key and query within each head. It is simple to convert any models with self-attention, including pre-trained ones, to the proposed alignment attention. On a variety of language understanding tasks, we show the effectiveness of our method in accuracy, uncertainty estimation, generalization across domains, and robustness to adversarial attacks.
  arXiv  Detail & Related papers  (2021-10-25T00:54:57Z)
• More Than Just Attention: Learning Cross-Modal Attentions with Contrastive Constraints [63.08768589044052]
  We propose Contrastive Content Re-sourcing ( CCR) and Contrastive Content Swapping ( CCS) constraints to address such limitation. CCR and CCS constraints supervise the training of attention models in a contrastive learning manner without requiring explicit attention annotations. Experiments on both Flickr30k and MS-COCO datasets demonstrate that integrating these attention constraints into two state-of-the-art attention-based models improves the model performance.
  arXiv  Detail & Related papers  (2021-05-20T08:48:10Z)

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.