How to Build a Model-Native Agent That Learns Internal Planning, Memory, and Multi-Tool Reasoning Through End-to-End Reinforcement Learning
"AI reasoning" 相关文章
北大字节开源首个时空推理视频模型，思考过程全透明，性能超越GPT-4o
Qwen“半成品”推理模型刷下AIME满分，俘获大批国外开发者！实测碾压GPT-5 Thinking、还能写侦探小说
开源模型首次物理奥赛IPhO夺金！上海AI Lab 235B模型击败GPT-5和Grok-4
Meta「透视」AI思维链：CRV推理诊断，准确率达 92%！
Meta「透视」AI思维链：CRV推理诊断，准确率达 92%！
R-HORIZON：长程推理时代来临，复旦NLP&美团LongCat重磅发布LRMs能力边界探测新范式
从会画画到会思考：快手可灵提出T2I-CoReBench，最强模型也难逃推理瓶颈
First Try Matters，不是Aha Moment：邴立东团队揭示推理模型靠首答，不靠反思
NeurIPS 2025 | 上交大提出MM-UPT：多模态大模型的“无监督后训练”范式
First Try Matters，不是Aha Moment：邴立东团队揭示推理模型靠首答，不靠反思
NeurIPS 2025 | 上交大提出MM-UPT：多模态大模型的“无监督后训练”范式
从会画画到会思考：快手可灵提出T2I-CoReBench，最强模型也难逃推理瓶颈
递归语言模型登场！MIT华人新作爆火，扩展模型上下文便宜又简单
AI Agents as Universal Task Solvers
CoT 之后，CoF 如何让帧间逻辑从「隐式对齐」变成「显式思考」？
2篇最新152页论文，RL+LLM被彻底讲透了！
Training Qwen-1.5B with a CoT legibility penalty
Tiny Recursive Model (TRM): A Tiny 7M Model that Surpass DeepSeek-R1, Gemini 2.5 pro, and o3-mini at Reasoning on both ARG-AGI 1 and ARC-AGI 2
Tiny Recursive Model (TRM): A Tiny 7M Model that Surpass DeepSeek-R1, Gemini 2.5 pro, and o3-mini at Reasoning on both ARG-AGI 1 and ARC-AGI 2
