0:000:00
<p>本期的 13 篇论文如下：</p><p>[00:25] 🧩 IterResearch: Rethinking Long-Horizon Agents via Markovian State Reconstruction（IterResearch：基于马尔可夫状态重构的长程智能体再思考）</p><p>[01:16] 🏆 DRIVE: Data Curation Best Practices for Reinforcement Learning with Verifiable Reward in Competitive Code Generation（DRIVE：面向可验证奖励强化学习的竞赛级代码生成数据精选最佳实践）</p><p>[02:03] 🔬 The Station: An Open-World Environment for AI-Driven Discovery（“站”：面向AI驱动科学发现的开放世界环境）</p><p>[02:43] 🚀 RedOne 2.0: Rethinking Domain-specific LLM Post-Training in Social Networking Services（RedOne 2.0：社交网络场景下领域大模型后训练新范式）</p><p>[03:15] 🧠 SofT-GRPO: Surpassing Discrete-Token LLM Reinforcement Learning via Gumbel-Reparameterized Soft-Thinking Policy Optimization（SofT-GRPO：用Gumbel重参数化软思考策略优化让离散Token强化学习望尘莫及）</p><p>[03:53] 🧭 Routing Manifold Alignment Improves Generalization of Mixture-of-Experts LLMs（路由流形对齐提升混合专家大语言模型的泛化能力）</p><p>[04:30] 🔍 Reasoning with Confidence: Efficient Verification of LLM Reasoning Steps via Uncertainty Heads（以置信度推理：通过不确定性头高效验证大模型推理步骤）...