评分依据：Establishes task-stratified scaling laws for quantized LLMs. Critical for understanding PTQ tradeoffs across capability types.
Task-Stratified Knowledge Scaling Laws for Post-Training Quantized Large Language Models
发布
采集
行业动态 7.0 分
— Establishes task-stratified scaling laws for quantized LLMs. Critical for understanding PTQ tradeoffs across capability types. 原文： arXiv