09Show HN: Reviving a 20-year-old puzzle game Chromatron with Ghidra and AIHACKER_NEWS Hacker News阅读文章 arrow_forward
08We should revisit literate programming in the agent eraHACKER_NEWS Hacker News AI Agent阅读文章 arrow_forward
08Agent Safehouse – macOS-native sandboxing for local agentsHACKER_NEWS Hacker News AI Agent阅读文章 arrow_forward
08Oracle may slash up to 30k jobs to fund AI data-centers as US banks retreatHACKER_NEWS Hacker News阅读文章 arrow_forward
08Phi-4-reasoning-vision and the lessons of training a multimodal reasoning modelHACKER_NEWS Hacker News阅读文章 arrow_forward
08SWE-CI: Evaluating Agent Capabilities in Maintaining Codebases via CIHACKER_NEWS Hacker News AI Agent阅读文章 arrow_forward
08Autoresearch: Agents researching on single-GPU nanochat training automaticallyHACKER_NEWS Hacker News AI Agent阅读文章 arrow_forward
07LLM Doesn't Write Correct Code. It Writes Plausible CodeHACKER_NEWS Hacker News 大语言模型阅读文章 arrow_forward
07Sarvam 105B, the first competitive Indian open source LLMHACKER_NEWS Hacker News 大语言模型阅读文章 arrow_forward
07LLMs work best when the user defines their acceptance criteria firstHACKER_NEWS Hacker News 大语言模型阅读文章 arrow_forward
06Towards Provably Unbiased LLM Judges via Bias-Bounded EvaluationARXIV ArXiv 大语言模型阅读文章 arrow_forward
06The Spike, the Sparse and the Sink: Anatomy of Massive Activations and Attention SinksARXIV ArXiv阅读文章 arrow_forward
06POET-X: Memory-efficient LLM Training by Scaling Orthogonal TransformationARXIV ArXiv 大语言模型阅读文章 arrow_forward
06Leveraging LLM Parametric Knowledge for Fact Checking without RetrievalARXIV ArXiv AI Agent阅读文章 arrow_forward
06Cheap Thrills: Effective Amortized Optimization Using Inexpensive LabelsARXIV ArXiv 机器学习阅读文章 arrow_forward
06Censored LLMs as a Natural Testbed for Secret Knowledge ElicitationARXIV ArXiv 大语言模型阅读文章 arrow_forward
06A tool that removes censorship from open-weight LLMsHACKER_NEWS Hacker News 大语言模型阅读文章 arrow_forward