Xinyu Li, Kexi Chen, Jiajie Shen Fudan University, Ying Zheng, Hong Lu Fudan University, Jin Zhao Fudan University, Xin Wang Fudan University
In multivariate long-term time series forecasting, it is widely believed that the effectiveness of self-attention arises from its attention matrix. We challenge this assumption with a counterintuitive finding: our experiments, conducted on three classic and three latest Transformer models, show that dot-product attention can be replaced by element-wise operations without token interaction, such as the addition and Hadamard product, while maintaining or even improving accuracy. This leads to our central hypothesis: the effectiveness of self-attention in this task stems not from the dynamic attention matrix, but from the multi-branch feature extraction inherent in the parallel projections to Query, Key, and Value matrices and their fusion. To validate this, we construct a simple multi-branch MLP that isolates the ‘multi-branch mapping with element-wise operation’ structure from the Transformer and show that it achieves competitive performance. Our results indicate that the source of performance in self-attention has been misattributed, suggesting that the true benefit lies in the architectural principle of multi-branch mapping and fusion, not in the attention matrix.
尚无 AI 解读，展示原文摘要
cs.CL
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表
Tianle Gu, Kexin Huang, Zongqi Wang, Yixu Wang, Jie Li Shanghai Artificial Intelligence Laboratory, Xin Wang Fudan University, Yang Yao, Yujiu Yang Tsinghua University 等 10 位作者
Safety alignment is a fundamental prerequisite for building trustworthy artificial general intelligence. Despite substantial progress in safety alignment techniques, empirical evidence shows that aligned large language models can still produce unsafe responses under minor internal perturbations, revealing a robustness gap in existing safety mechanisms at the latent representation level. In this paper, we study the robustness evaluation of safety alignment under latent-space perturbations. We introduce Activation Steering Attack (ASA), and leverage the Negative Log-Likelihood (NLL) as a diagnostic signal to probe the local sensitivity of safety behaviors in latent space. By measuring a model’s likelihood under controlled perturbations to its hidden representations, we assess the stability of its original responses. The probing signal is model-agnostic and supervision-free, enabling a general and reproducible diagnostic metric for analyzing safety robustness. Leveraging these probes, we systematically uncover a set of previously underexplored empirical findings, including (1) non-stationarity of layer vulnerabilities, revealing that the most vulnerable layer is an unstable property and even relocates after robustness training; (2) instance-level alignment with cross-layer consistency, where specific inputs remain universally vulnerable across the entire model hierarchy; (3) compositional effects of ASA, characterized by its incremental accumulation across sequential decoding steps and its potential for prompt-level jailbreak effectiveness.
尚无 AI 解读，展示原文摘要
cs.CL
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表