Recent Posts

On the Off-Policy Teacher in On-Policy Distillation

同策略蒸馏中的离策略教师困局

10/3/2026

arxivaics.ai

Semifactual Credit-Augmented Policy Optimization

半事实信用增强策略优化：瓦解 GRPO 的虚假关联

10/1/2026

arxivaics.ai

Imagine3D-LLM: Teaching MLLMs to Imagine 3D Scenes Before Answering

Imagine3D-LLM：让多模态大模型在作答前先构想 3D 场景

9/30/2026

arxivaics.cl

Telescopic Language Models

望远镜语言模型：单次训练贯通全深度算力连续谱

9/29/2026

arxivaics.ai

Learning to Stop without Learning to Stop: Self-Supervised Confidence Training Improves Reasoning Efficiency

无心插柳的收敛：自监督置信度训练如何自发提升大模型推理效率

9/28/2026

arxivaics.ai

LLM Agents Can Easily Tamper With Their Own Traces

大模型智能体能轻易篡改自身执行轨迹

9/25/2026

arxivaics.cr

On the Diffusibility of High-Dimensional Latents

高维表征潜空间的扩散可行性之谜

9/24/2026

arxivaics.cv

Flash-dLLM: IO-Aware KV Caching and Parallel Decoding for Fast Diffusion LLMs

Flash-dLLM：IO 感知 KV 缓存与并行解码突破扩散大模型推理瓶颈

9/23/2026

arxivaics.cl

Critical-State RL: Diagnosing Trainable States for Multi-Turn Tool Use

关键状态强化学习：诊断多轮工具调用的可训练节点

9/22/2026

arxivaics.cl

Designer-RSI: Evolving Procedural Memory from User Traffic for Agentic Graphic Design

Designer-RSI：从真实用户交互中进化程序性设计记忆

9/21/2026

arxivaics.ai

An Empirical Study of Harness Design for Coding Agents

编码智能体 Harness 设计的实证剖析

9/18/2026

arxivaics.ai

dQwen3.5: Hybrid-Attention Diffusion Language Models

dQwen3.5：混合注意力架构突围扩散语言模型

9/18/2026

arxivaics.cl