📄 中文摘要
一篇新发布的arXiv论文改变了我们对人工智能推理的理解。研究人员发现，在“自组织临界性”（SOC）条件下训练的大型语言模型（LLM）会自发发展出推理能力，这一现象类似于物理学中的相变。这一发现表明，LLM在特定条件下能够展现出超出传统预期的智能行为。用户无需等待实验室的产品化，可以通过NexaAPI访问市场上最强大的推理LLM，享受最具性价比的人工智能推理API服务。
PLDR-LLM：人人都在谈论的人工智能推理突破（附免费API教程）
📄 English Summary
PLDR-LLM: The AI Reasoning Breakthrough Everyone Is Talking About (+ Free API Tutorial)
A newly released paper on arXiv is reshaping the understanding of AI reasoning. Researchers discovered that large language models (LLMs) trained under conditions of 'self-organized criticality' (SOC) spontaneously develop reasoning abilities, akin to phase transitions in physics. This finding indicates that LLMs can exhibit intelligence beyond traditional expectations under specific conditions. Users do not need to wait for labs to productize this technology; they can access the most capable reasoning LLMs today through NexaAPI, which offers the most cost-effective AI inference API on the market.