English

On Zero-Initialized Attention: Optimal Prompt and Gating Factor Estimation

Machine Learning 2025-06-19 v3

Abstract

The LLaMA-Adapter has recently emerged as an efficient fine-tuning technique for LLaMA models, leveraging zero-initialized attention to stabilize training and enhance performance. However, despite its empirical success, the theoretical foundations of zero-initialized attention remain largely unexplored. In this paper, we provide a rigorous theoretical analysis, establishing a connection between zero-initialized attention and mixture-of-expert models. We prove that both linear and non-linear prompts, along with gating functions, can be optimally estimated, with non-linear prompts offering greater flexibility for future applications. Empirically, we validate our findings on the open LLM benchmarks, demonstrating that non-linear prompts outperform linear ones. Notably, even with limited training data, both prompt types consistently surpass vanilla attention, highlighting the robustness and adaptability of zero-initialized attention.

Keywords

Cite

@article{arxiv.2502.03029,
  title  = {On Zero-Initialized Attention: Optimal Prompt and Gating Factor Estimation},
  author = {Nghiem T. Diep and Huy Nguyen and Chau Nguyen and Minh Le and Duy M. H. Nguyen and Daniel Sonntag and Mathias Niepert and Nhat Ho},
  journal= {arXiv preprint arXiv:2502.03029},
  year   = {2025}
}

Comments

Accepted at ICML 2025

R2 v1 2026-06-28T21:33:14.152Z