Language as a Wave Phenomenon: Semantic Phase Locking and Interference in Neural Networks

In standard Transformer architectures, semantic importance is often conflated with activation magnitude, obscuring the geometric structure of latent representations. To disentangle these factors, we introduce PRISM, a complex-valued architecture designed to isolate the computational role of phase. By enforcing a strict unit-norm constraint (|z| = 1) and replacing attention with gated harmonic convolutions, the model is encouraged to utilize subtractive interference in the frequency domain to suppress noise, rather than relying on magnitude-based gating. We utilize this constrained regime to study a hybrid architecture—fusing phase-based routing with standard attention—which achieves improved parameter efficiency and representation quality compared to baselines in our evaluated settings. Mechanistically, interventional ablations indicate that the model carries substantial task-relevant information in phase: preserving phase largely maintains performance, whereas disrupting phase causes severe degradation. Together, these results suggest that phase-based spectral interference is a usable computational mechanism for neural sequence modeling at the evaluated scale.

Modern language models usually represent information as numbers, where larger values often matter more. This paper studies whether complex-valued models can instead use phase, the angle of a representation, to carry information. We introduce PRISM to test this idea in sequence modeling. Experiments show that preserving phase keeps most performance, while disrupting phase causes severe degradation. This suggests that phase can carry meaningful relational information and may help design more efficient sequence models.