פיתוח כולל-הפוך-לפי-עומק מתוך לא-ריליניאריות בתנאי התייחסות

תקציר מקורי באנגליתarXiv:2610.11063v1 Announce Type: new Abstract: Scaling laws describe power-law improvements in model performance with dataset size and parameter count, yet their underlying mechanisms are not fully understood. To explain the parameter count scaling, existing theory posits power-law scaling with model depth. In linear-attention models, this scaling is tied to a power-law data spectrum: unable to selectively attend to relevant tokens, these models learn according to global spectral strength, with stronger directions learned before weaker ones. Large language models, however, can be strongly nonlinear. Here, we show that nonlinear attention yields inverse-depth decay of loss across all tested data spectra. Nonlinearity enables attention to focus selectively on relevant tokens, allowing stron