BiSpikCLM: A Spiking Language Model integrating Softmax-Free Spiking Attention and Spike-Aware Alignment Distillation

ArXi:2605.13859v1 Announce Type: cross Spiking Neural Networks (SNNs) offer promising energy-efficient alternatives to large language models (LLMs) due to their event-driven nature and ultra-low power consumption. However, to preserve capacity, most existing spiking LLMs still incur intensive floating-point matrix multiplication (MatMul) and nonlinearities, or