只训练10%关键注意力头，就能提升大模型对齐效果
ALPS: Attention Localization and Pruning Strategy for Efficient Alignment of Large Language Models
- 定位任务敏感注意力头，仅对其更新以降低调参成本
- 仅激活10%参数仍比基线高2%性能，三任务验证有效
- 识别的注意力头可跨数据集复用，防止知识遗忘
将通用大语言模型适配到下游任务通常需要高昂的训练开销。现有方法多依赖少量数据或数据驱动激活关键注意力头，但存在数据依赖问题，影响泛化性与复用性。为此，我们提出注意力定位与剪枝策略（ALPS），通过定位最敏感的任务相关注意力头，并限制训练时仅更新这些头，从而显著降低对齐成本。实验表明，该方法在微调过程中仅激活10%的注意力参数，却在三个任务上实现比基线高出2%的性能。此外，所识别出的任务特异性注意力头具备跨数据集迁移能力，有效缓解知识遗忘问题。本研究为高效大模型对齐提供了新视角。代码已开源：https://github.com/VoiceBeer/ALPS。
原文摘要 · Abstract (English)
Aligning general-purpose large language models (LLMs) to downstream tasks often incurs significant training adjustment costs. Prior research has explored various avenues to enhance alignment efficiency, primarily through minimal-data training or data-driven activations to identify key attention heads. However, these approaches inherently introduce data dependency, which hinders generalization and reusability. To address this issue and enhance model alignment efficiency, we propose the Attention Localization and Pruning Strategy (ALPS), an efficient algorithm that localizes the most task-sensitive attention heads and prunes by restricting attention training updates to these heads, thereby reducing alignment costs. Experimental results demonstrate that our method activates only 10% of attention parameters during fine-tuning while achieving a 2% performance improvement over baselines on three tasks. Moreover, the identified task-specific heads are transferable across datasets and mitigate knowledge forgetting. Our work and findings provide a novel perspective on efficient LLM alignment. The code is available at https://github.com/VoiceBeer/ALPS.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。