医疗视觉问答也能“知道何时不懂”，置信度感知推理让AI诊断更可靠
arXiv:2608.10964v1 Announce Type: new Abstract: Reinforcement Fine-Tuning (RFT) has enabled medical Multimodal Large Language Models (MLLMs) to produc…
开源大模型以细粒度辩论实现可信数据扩充，为心理健康与网络内容安全标注开辟新路径。
arXiv:2512.06227v3 Announce Type: replace-cross Abstract: Real-world indicators play an important role in many Natural Language Processing (NLP) appli…
开源授权引擎Lelu，专为AI代理设计，支持置信度感知和策略驱动，从个人项目到生产集群都能用。
Article URL: https://lelu-ai.com Comments URL: https://news.ycombinator.com/item?id=48299318 Points: 1 # Comments: 0
利用解码步骤的置信度信号，提出混合精度存储的KV缓存逐出策略，有效缓解长序列推理中的GPU内存瓶颈。
arXiv:2605.24786v1 Announce Type: cross Abstract: Long-horizon LLM inference turns the key--value (KV) cache into the dominant GPU memory consumer and…