Scaling recommendation models is a central challenge in recommender systems. Recently, RankMixer has emerged as an effective solution, operating on a unified token representation and alternating between token mixing and per-token feedforward networks (P-FFNs) to achieve scalable performance. However, RankMixer suffers from \textit{embedding collapse}, where learned representations have low effective rank, limiting expressivity and underutilizing the expanded representation space. Through empirical analysis and theoretical insights, we identify rigid token mixing and P-FFN modules as the primary causes of this phenomenon, jointly inducing a \textbf{damped oscillatory trajectory} in effective-rank evolution across layers. To address it, we propose RankElastor, a novel architecture that produces spectrum-robust representations with provable collapse mitigation. RankElastor introduces two components: (i) \textbf{parameterized full mixing}, which enables expressive token mixing with improved spectral robustness; and (ii) \textbf{GLU-improved P-FFNs}, which stabilize representation spectra through GLU-style FFN modules. Extensive experiments on large-scale industrial datasets demonstrate that RankElastor consistently improves recommendation performance, mitigates embedding collapse, and exhibits robust scaling behavior. Code is available at this GitHub repository: https://github.com/vasile-paskardlgm/RankElastor

翻译：扩展推荐模型是推荐系统中的核心挑战。近期，RankMixer作为一种有效方案出现，它基于统一的令牌表示进行操作，交替进行令牌混合和每令牌前馈网络（P-FFN），以实现可扩展性能。然而，RankMixer存在\textit{嵌入坍塌}问题，即学习到的表示具有低有效秩，限制了表达能力且未能充分利用扩展后的表示空间。通过实证分析和理论洞察，我们确定刚性令牌混合和P-FFN模块是这一现象的主因，它们共同导致跨层有效秩演化呈现\textbf{阻尼振荡轨迹}。为解决该问题，我们提出RankElastor——一种新型架构，可生成频谱鲁棒的表示并具有可证明的坍塌缓解能力。RankElastor引入两个组件：(i) \textbf{参数化全混合}，通过改进的频谱鲁棒性实现表达性令牌混合；(ii) \textbf{GLU改进型P-FFN}，通过GLU风格的FFN模块稳定表示频谱。在大规模工业数据集上的广泛实验表明，RankElastor能持续提升推荐性能、缓解嵌入坍塌，并展现出稳健的扩展行为。代码已开源：https://github.com/vasile-paskardlgm/RankElastor