Latest

MetaCtrl addresses a critical inefficiency in reasoning models by dynamically determining when additional computation is beneficial, eliminating the need for predefined budgets or retraining. This plug-and-play solution significantly improves accuracy while reducing inference costs, making it valuable for developers working with resource-constrained environments. The open-source implementation allows immediate integration into existing reasoning pipelines.

#MetaCtrl #DeepSeek-R1-Distill-Qwen-7B #Qwen3-14B #reasoning models

arXiv: DeepSeek@Zhibin Wen et al. (5)Source

Models BuildersSelected09:37

EASE framework for self-evolving agents

EASE addresses a critical limitation in agent skill curation by solving cross-executor degradation through behavior adaptation. Developers working with multiple AI models can now use a single, more efficient curator that reduces inference tokens by 9.1-14.5% while maintaining performance across diverse executors, making it particularly valuable for organizations deploying multi-model systems.

#EASE #skill curation #self-evolving agents #Qwen3

Event

arXiv: DeepSeek@Zhen Xiong, Qiaoyu Tan2 sourcesSource

9月18日

Models US vs China11:52

Chinese content智谱发布混合注意力扩散语言模型dQwen3.5系列

智谱新出的dQwen3.5系列模型，用混合注意力架构改造自Qwen3.5，在训练效率上比全注意力架构有优势。

#dQwen3.5 #智谱 #混合注意力 #扩散模型

arXiv cs.LG@Anton Xue et al. (6)Source

Models BuildersSelected11:21

ActObs changes how agents explore under RL

This research provides a new method for improving the exploration capabilities of AI agents, which could lead to more robust and effective systems in various applications.

#ActObs #Qwen3-4B #reinforcement learning #terminal-bench

arXiv cs.AI@Juzheng Zhang et al. (5)Source

9月17日

AITOP16:09

Anthropic拟10月在新加坡设亚太第五个办事处开...

Anthropic拟10月在新加坡设亚太第五个办事处开放本地招聘数据显示，按人口比例计算，新加坡的Claude.ai使用水平在121个国家中排名第二。我们观察到，人工智能公司Anthropic计划于10月在新加坡设立其亚太地区的第五…

AITOPRead

9月15日

Models BuildersSelected13:06

Physically Partitioned KVCache Format for MoE Inference

This research provides a method to improve the performance of large language models on GPUs by better managing memory usage, which is relevant for anyone working with or evaluating MoE models.

#ai-models #Mixture-of-Experts #MoE #inference

arXiv: DeepSeek@Enda Yu et al. (3)Source

9月9日

Models BuildersSelected73°11:15

PlannerForge Framework for Autonomous Testing

PlannerForge improves autonomous driving testing efficiency, with open-source models matching commercial performance and reducing collisions by over 50% through cost-tuning.

#PlannerForge #Qwen3.6:35B #autonomous driving #LLM agents

arXiv cs.AI@Yuan Gao et al. (8)Source

9月7日

Models BuildersSelected73°09:23

ACE framework for MoE-based LLMs

ACE improves efficiency for MoE models without retraining, benefiting developers working with large language models in resource-constrained environments.

#ACE #MoE #Qwen3.6-35B-A3B #expert skipping

arXiv cs.AI@Zukang Xu et al. (8)Source

9月4日

ModelsSelected73°11:25

Terminal-Universe Framework for Agent Environments

Developers can use Terminal-Universe to create more realistic training environments for terminal-based code agents, improving their performance on real-world coding tasks.

#Terminal-Universe #Qwen3.5-27B #Terminal-Bench #code agents

arXiv cs.AI@Jie Wu et al. (14)Source

9月2日

ModelsSelected73°09:36

HiveTraceGuard-Pro generative guardrail model

Released under Apache-2.0 license with lowest median latency at 14.3 ms, offering robust protection against Russian and English prompt injections.

#HiveTraceGuard-Pro #Qwen3-0.6B #guardrail models #prompt injection

arXiv cs.AI@Nikita Oblakov et al. (3)Source

9月1日

Models78°10:14

Chinese contentTuringLLM推出20B参数专家混合模型

Turing团队发布20B参数专家混合模型，每激活2B参数，性能接近9B模型，支持超长上下文。

#TuringLLM #Turing-20B-A2B #专家混合模型 #长上下文

arXiv cs.AI@Yuheng Zhang et al. (22)Source

8月27日

ModelsMulti-sourceSelected09:21

JIT-Agent: Harness Intelligence Model

JIT-Agent improves performance across multiple model families including DeepSeek V4 and Qwen3.6, offering a scalable alternative to manual harness design.

#JIT-Agent #DeepSeek-V4-Flash #GLM-5.2 #agents

Event

arXiv: DeepSeek@Guibin Zhang et al. (16)5 sourcesSource

8月20日

Models US vs China10:18

Chinese contentQwen3模型推出VLA框架监测多代理隐蔽通信

Qwen3推出了VLA框架，用来监测多代理潜在通信，在小范围拍卖里表现好，你可以去了解下。

#Qwen3 #智能体 #多模态 #推理模型

arXiv cs.AI@Ramneet Kaur et al. (6)Source

8月19日

ModelsSelected11:54

Chinese contentCABLE：通过互补前导链接扩展记忆检索范围