2025年4月14日多模态大模型论文推送

文章预览

  LLM http://arxiv.org/pdf/2504.08672v1 标题： Genius: A Generalizable and Purely Unsupervised Self-Training Framework For Advanced Reasoning 关键词：unsupervised learning, self-training, reasoning 简介：上海AI LAb等提出了Genius，一个基于无监督的self-training框架，用于提升LLM的reasoning能力 http://arxiv.org/pdf/2504.08690v1 标题： Fast-Slow-Thinking: Complex Task Solving with Large Language Models 关键词：Task Decomposition, Complex Problem Solving 简介：南京大学等提出了Fast-Slow-Thinking (FST)，一个提升LLM解决复杂任务能力的方法 http://arxiv.org/pdf/2504.07866v2 标题：  Pangu Ultra: Pushing the Limits of Dense Large Language Models on Ascend NPUs 关键词：NPU, LLM 简介：华为发布了Pangu Ultra的技术报告，Pangu Ultra是在NPU上训练的，有135B参数的LLM http://arxiv.org/pdf/2504.08590v1 标题： Playpen: An Environment for Exploring Learning Through Conversational Interaction 关键词：synt ………………………………