3 months agoAI News CN (Telegram)#Update #OpenAI #ChatGPT GPT-4.5 现已发布： - OpenAI目前为止最大的模型 - 宣称幻觉现象低于 GPT-4o、o1、o3-mini - 宣称无需推理即可具备一定的理解能力和创...
3 months agoAI News CN (Telegram)🖼 OpenAI发布GPT-4.5预览版 GPT-4.5是OpenAI目前除o3mini以外最强大的模型，通过扩展非监督学习，提升了模式识别和知识理解能力。早期测试表明，GPT-4.5在遵...
3 months agoHugging Face Daily PapersMobius: Text to Seamless Looping Video Generation via Latent Shift
3 months agoHugging Face Daily PapersFlexiDiT: Your Diffusion Transformer Can Easily Generate High-Quality Samples with Less Compute
3 months agoHugging Face Daily PapersR1-T1: Fully Incentivizing Translation Capability in LLMs via Reasoning Learning
3 months agoHugging Face Daily PapersUniTok: A Unified Tokenizer for Visual Generation and Understanding
3 months agoHugging Face Daily PapersCODESYNC: Synchronizing Large Language Models with Dynamic Code Evolution at Scale
3 months agoHugging Face Daily PapersSoRFT: Issue Resolving with Subtask-oriented Reinforced Fine-Tuning
3 months agoHugging Face Daily PapersR2-T2: Re-Routing in Test-Time for Multimodal Mixture-of-Experts
3 months agoAI News CN (Telegram)OpenAI 的 GPT-4.5 更善于说服其他AI给它打钱 本周四，OpenAI 发布了一份白皮书，描述了其代号为 Orion 的 GPT-4.5 模型的能力。 根据该论文，OpenAI 对该模型进...
3 months agoAI News CN (Telegram)OpenAI 发布 GPT-4.5 迄今为止规模最大、知识最丰富的模型 基准测试数据表明，GPT-4.5 比 GPT-4o 的升级幅度不大。 在 SWE-bench Verified 基准测试中，GPT-4.5 ...
3 months agoHugging Face Daily PapersAdapting Automatic Speech Recognition for Accented Air Traffic Control Communications
3 months agoHugging Face Daily PapersMMKE-Bench: A Multimodal Editing Benchmark for Diverse Visual Knowledge
3 months agoAI News CN (Telegram)Elevated errors for requests to Claude 3.5 Sonnet 0620 Feb 27, 11:51 PST Identified - The issue has been identified and a fix is being implemented. vi...
3 months agoAI News CN (Telegram)较年前涨价10万 H20已成“大厂最爱” 事实也确是如此。得益于数据中心业务，英伟达2025第四财季交出了一份亮眼的成绩单。算力资源信息共享平台柏林云负责人告诉....