Skip to content
arXiv cs.CL · Papers

Preserving Fairness and Safety in Quantized LLMs Through Critical Weight Protection

arXiv:2601.12033v2 Announce Type: replace Abstract: Quantization is widely adopted to reduce the computational cost of large language models (LLMs); however, its implications for fairness and safety, particularly in dynamic quantization and multilingual contexts, remain underexplored. In this work, we conduct a systema