Curating the frontier of Artificial Intelligence

Procedural Graphs: Self-Evolving Execution Structures for LLM Agents

大型语言模型越来越多地被部署为能够进行长期规划并通过外部工具执行操作的代理。大多数代理选择行动通过...

Large language models are increasingly deployed as agents that plan over long horizons and act through external tools. Most agents select actions thro...

Don't Drop Dropout: Optimizing Layer Sparsity for Efficient LLM Training and Inference

Layer dropout (a.k.a.随机深度) 已被证明可以实现更快的训练, 更高的准确性, 以及对零样本层剪枝的鲁棒性...

Layer dropout (a.k.a. stochastic depth) has been shown to enable faster training, higher accuracy, and robustness to zero-shot layer pruning in both l...

Testing Interchangeability in LLM Agent Teams

Production multi-agent systems replace agents constantly, on the assumption that an agent filling a role is interchangeable with any other agent that ...

Production multi-agent systems replace agents constantly, on the assumption that an agent filling a role is interchangeable with any other agent that ...

GUT: Quantifying and Optimizing the Reasoning Uncertainty of LLMs via Graph Complexity

近年来，大型语言模型(LLMs) 的推理能力取得了巨大进步。然而, LLM的推理过程通常...

Recent years have witnessed great advances in the reasoning ability of Large Language Models (LLMs). However, the reasoning processes of LLMs often ex...

Beyond Aggregate Scores: Behavioral Correctness Assumptions for Assessing Reference-Based Automatic Evaluation Methods

基于参考的自动评估方法在评估自然语言生成系统中发挥着至关重要的作用。现有的元评估主要...

Automated reference-based evaluation methods play a critical role in assessing natural language generation systems. Existing meta-evaluation primarily...

RISE: Recursive Improvement via Self-Extrapolating Policy Distillation

在策略蒸馏(OPD) 为语言模型训练后, 提供密集的, 每个标记监督，但其有效性受到教师的瓶颈......

On-policy distillation (OPD) provides dense, per-token supervision for language model post-training, but its effectiveness is bottlenecked by teacher ...

Large Language Models for HVAC Operations in Building Energy Systems: A Critical Review of Methods, Applications, and Deployment Readiness

楼宇自动化系统生成丰富的传感器数据，但仍然缺乏洞察力，因为异构点命名, 缺少元数据, 和碎片化文档...

Building automation systems generate rich sensor data yet remain insight-poor because heterogeneous point naming, missing metadata, and fragmented doc...

LLM-Driven Algorithm Design for Quantum Circuit Synthesis based on Binary Decision Diagrams

量子电路是在量子器件, 上实现量子算法的核心，其中量子门必须是可逆的。许多量子算法...

Quantum circuits are central to implementing quantum algorithms on quantum devices, where quantum gates must be reversible. Many quantum algorithms re...

Technical Manual for a Toolkit for Measuring Contextual Individuation in Transformer Language Models

Transformer 语言模型将单个 , 上下文无关向量分配给其嵌入层 , 的单词类型，但人们普遍认为它可以个性化......

A transformer language model assigns a single, context-independent vector to a word type at its embedding layer, yet is widely believed to individuate...

Does Your Agent's Memory Survive a Model Upgrade? A Controlled Study of Memory Portability

模型升级是例行公事，而 3B 内存迁移则不然。代理可以保留相同的内存存储，但仍然会忘记:新模型可能会解释旧的注释......

Model upgrades are routine; memory migrations are not. An agent can keep the same memory store and still forget: a new model may interpret old notes d...