1
BitsMoE: Efficient Spectral Energy-Guided Bit Allocation for MoE LLM Quantization
混合专家模型量化新方法：基于频谱能量引导的自适应比特分配，显著提升压缩效率与模型性能。
arXiv:2606.00079v1 Announce Type: new Abstract: Mixture-of-Experts (MoE) large language models reduce per-token computation through sparse expert acti…