今天的五篇论文展示了AI在推理、效率和理解上的突破: Learning to Reason under Off-Policy Guidance:提出LUFFY框架,通过外部高质量推理轨迹提升AI数学推理能力,泛化性强,平均提升7%。 Efficient Pretraining Length Scaling:PHD-Transformer通过智能管理KV缓存,实现高效预训练长度缩放,性能提升1.5%-2%,推理速度几乎不变。 MAGIC: Near-Optimal Data Attribution for Deep Learning:MAGIC方法精准预测训练数据对模型的影响,相关性高达0.96,助力模型调试和可解释性。 Exploring How LLMs Capture and Represent Domain-Specific Knowledge:发现大型语言模型在预填充阶段形成领域特定轨迹,用于智能路由,准确率提升12.3%。 LongPerceptualThoughts: Distilling System-2 Reasoning for System-1 Perception:通过合成复杂推理数据,视觉-语言模型学会深度推理,视觉任务提升3.4分,文本推理也意外提升。完整推介:https://mp.weixin.qq.com/s/0IlcYwqQ-GAgZDgh5TCnNQ
No persons identified in this episode.
This episode hasn't been transcribed yet
Help us prioritize this episode for transcription by upvoting it.
Popular episodes get transcribed faster
Other recent transcribed episodes
Transcribed and ready to explore now
SpaceX Said to Pursue 2026 IPO
10 Dec 2025
Bloomberg Tech
Don’t Call It a Comeback
10 Dec 2025
Motley Fool Money
Japan Claims AGI, Pentagon Adopts Gemini, and MIT Designs New Medicines
10 Dec 2025
The Daily AI Show
Eric Larsen on the emergence and potential of AI in healthcare
10 Dec 2025
McKinsey on Healthcare
What it will take for AI to scale (energy, compute, talent)
10 Dec 2025
Azeem Azhar's Exponential View
Reducing Burnout and Boosting Revenue in ASCs
10 Dec 2025
Becker’s Healthcare -- Spine and Orthopedic Podcast