Plug-and-Play Spiking Operators: Breaking the Nonlinearity Bottleneck in Spiking Transformers
For researchers deploying spiking neural networks for large language models, this work provides a practical solution to integrate nonlinearities into neuromorphic hardware constraints.
The paper addresses the lack of spike-friendly implementations for key nonlinear operators in ANN-to-SNN conversion for Transformers, proposing a plug-and-play framework that approximates Softmax, SiLU, and normalization using LIF neuron groups and bit-shift scaling. The method achieves less than 1% accuracy drop across evaluated LLM tasks without fine-tuning.
ANN-to-SNN conversion offers a practical, training-free route to spiking large language models. However, current pipelines primarily focus on spike-driven realizations for Transformer linear-algebra operations, while providing limited support for key nonlinear operators. This gap limits compatibility with neuromorphic-style execution constraints, where such nonlinearities typically require division, exponentiation, or norm computations that are not naturally supported by standard leaky integrate-and-fire dynamics. To solve this problem, we propose a plug-and-play framework that implements spike-friendly approximations for Transformer nonlinearities and integrates into existing ANN-to-SNN pipelines. Our method decomposes these nonlinear computations into three recurring primitives -- division, exponentiation, and $\ell_2$ norms -- and realizes them via population computation using LIF neuron groups, combined with lightweight bit-shift scaling to avoid floating-point arithmetic. By composing these primitives as modular operator blocks, our framework supports common Transformer nonlinearities (e.g., Softmax, SiLU, and normalization) without any fine-tuning. Experiments on a range of LLMs Transformers show that selectively replacing the targeted nonlinear operators incurs less than a $1\%$ accuracy drop across all evaluated tasks.