LG - 机器学习 CV - 计算机视觉 CL - 计算与语言 AS - 音频与语音 RO - 机器人1、[CL] How Do Large Language Models Acquire Factual Knowledge During Pretraining?2、[CL] Creativity Has Left the Chat：The Price of Debiasing Language Models3、[LG] ExCP：Extreme LLM Checkpoint Compression via Weight-Momentum Joint Shrinking4、[LG] A Label is Worth a Thousand Images in Dataset Distillation5[CV] Task Me Anything摘要：大型语言模型如何在预训练过程中获取事实知识、大语言模型对齐的代价、通过权重-动量联合收缩实现极端LLM检查点压缩、软标签中结构化语义信息在数据集蒸馏中的作用至关重要、基于用户需求的基准测试生成引擎1、[CL] How Do Large Language Models Acquire Factual Knowledge During Pretraining?H Chang, J Park, S Ye, S Yang…[KAIST & UCL]大型语言模型如何在预训练过程中获取事实知识？要点:本研究通过向OLMo的中间检查点注入虚构知识并测量记忆
………………………………
