arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2602.22591cs.IR
收藏

相关性如何出现:对零样本重排序内部注意力的分层研究

Where Relevance Emerges: A Layer-Wise Study of Internal Attention for Zero-Shot Re-Ranking

The University of Queensland(昆士兰大学)

Haodong Chen, Shengyao Zhuang, Zheng Yao, Guido Zuccon, Teerapong Leelanupab

AI总结:

本文研究了多排名框架中生成、似然和内部注意力机制的对比,发现Transformer层中相关性信号呈现钟形分布,提出Selective-ICR策略降低推理延迟30%-50%,并在BRIGHT基准测试中证明了内部信号在复杂推理排序中的潜力。

摘要:

零样本文档重排序使用大语言模型(LLMs)已从点wise方法发展到listwise和setwise方法,以优化计算效率。尽管这些方法成功,但它们主要依赖生成评分或输出logits,这在推理延迟和结果一致性上存在瓶颈。在上下文重排序(ICR)最近被提出作为O(1)的替代方法。ICR直接提取内部注意力信号,避免文本生成的开销。然而,现有ICR方法仅在所有层上聚合信号;层间贡献及其在不同架构中的一致性尚未被探索。此外,没有统一的研究在一致条件下比较内部注意力与传统生成和似然机制在多种排名框架中的表现。在本文中,我们对生成、似然和内部注意力机制进行了正交评估,跨多个排名框架。我们进一步发现,相关性信号在Transformer层中呈现普遍的“钟形”分布,这促使提出了Selective-ICR策略,通过降低推理延迟30%-50%而不牺牲有效性。最后,在推理密集型BRIGHT基准测试中评估显示,精确捕捉高质量的上下文注意力信号从根本上减少了对模型扩展和强化学习的需求:零样本8B模型表现与14B强化学习重排序器相当,而甚至0.6B模型也优于最先进的生成方法。这些发现重新定义了基于LLM的重排序效率-效果前沿,并突显了内部信号在复杂推理排序任务中的潜在价值。我们的代码和结果可在https://github.com/ielab/Selective-ICR公开获取。

英文摘要:

Zero-shot document re-ranking with Large Language Models (LLMs) has evolved from Pointwise methods to Listwise and Setwise approaches that optimize computational efficiency. Despite their success, these methods predominantly rely on generative scoring or output logits, which face bottlenecks in inference latency and result consistency. In-Context Re-ranking (ICR) has recently been proposed as an O(1) alternative method. ICR extracts internal attention signals directly, avoiding the overhead of text generation. However, existing ICR methods simply aggregate signals across all layers; layer-wise contributions and their consistency across architectures have been left unexplored. Furthermore, no unified study has compared internal attention with traditional generative and likelihood-based mechanisms across diverse ranking frameworks under consistent conditions. In this paper, we conduct an orthogonal evaluation of generation, likelihood, and internal attention mechanisms across multiple ranking frameworks. We further identify a universal "bell-curve" distribution of relevance signals across transformer layers, which motivates the proposed Selective-ICR strategy that reduces inference latency by 30%-50% without compromising effectiveness. Finally, evaluation on the reasoning-intensive BRIGHT benchmark shows that precisely capturing high-quality in-context attention signals fundamentally reduces the need for model scaling and reinforcement learning: a zero-shot 8B model matches the performance of 14B reinforcement-learned re-rankers, while even a 0.6B model outperforms state-of-the-art generation-based approaches. These findings redefine the efficiency-effectiveness frontier for LLM-based re-ranking and highlight the latent potential of internal signals for complex reasoning ranking tasks. Our code and results are publicly available at https://github.com/ielab/Selective-ICR.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑