Posts

All the articles I've posted.

6.0
MemEvoBench: Benchmarking Memory MisEvolution in LLM Agents
2026年04月20日
· cs.CL updates on arXiv.org· 04/20 17:04 采集
AI领域重要新闻和深度分析
6.0
AI Agents and Hard Choices
2026年04月20日
· cs.AI updates on arXiv.org· 04/20 17:04 采集
AI领域重要新闻和深度分析
7.0
GTA-2: Benchmarking General Tool Agents from Atomic Tool-Use to Open-Ended Workflows
2026年04月20日
· cs.CL updates on arXiv.org· 04/20 17:04 采集
AI领域重要新闻和深度分析
7.0
LLMs Corrupt Your Documents When You Delegate
2026年04月20日
· cs.CL updates on arXiv.org· 04/20 17:04 采集
AI领域重要新闻和深度分析
6.0
Chain-of-Thought Degrades Visual Spatial Reasoning Capabilities of Multimodal LLMs
2026年04月20日
· cs.AI updates on arXiv.org· 04/20 17:04 采集
AI领域重要新闻和深度分析
6.0
Symbolic Guardrails for Domain-Specific Agents: Stronger Safety and Security Guarantees Without Sacrificing Utility
2026年04月20日
· cs.AI updates on arXiv.org· 04/20 17:04 采集
AI领域重要新闻和深度分析
6.0
The Semi-Executable Stack: Agentic Software Engineering and the Expanding Scope of SE
2026年04月20日
· cs.AI updates on arXiv.org· 04/20 17:04 采集
AI领域重要新闻和深度分析
6.0
HarmfulSkillBench: How Do Harmful Skills Weaponize Your Agents?
2026年04月20日
· cs.AI updates on arXiv.org· 04/20 17:04 采集
AI领域重要新闻和深度分析
7.0
Subliminal Transfer of Unsafe Behaviors in AI Agent Distillation
2026年04月20日
· 04/20 12:00 采集
7.0
LLM Reasoning Is Latent, Not the Chain of Thought
2026年04月20日
· 04/20 12:00 采集
7.0
MEDLEY-BENCH: Scale Buys Evaluation but Not Control in AI Metacognition
2026年04月20日
· 04/20 12:00 采集
6.5
LinuxArena: A Control Setting for AI Agents in Live Production Software Environments
2026年04月20日
· 04/20 12:00 采集
6.5
LACE: Lattice Attention for Cross-thread Exploration
2026年04月20日
· 04/20 12:00 采集
6.0
The World Leaks the Future: Harness Evolution for Future Prediction Agents
2026年04月20日
· 04/20 12:00 采集
6.0
Bilevel Optimization of Agent Skills via Monte Carlo Tree Search
2026年04月20日
· 04/20 12:00 采集
6.0
ReactBench: A Benchmark for Topological Reasoning in MLLMs on Chemical Reaction Diagrams
2026年04月20日
· 04/20 12:00 采集
6.0
Integrating Graphs, Large Language Models, and Agents: Reasoning and Retrieval
2026年04月20日
· 04/20 12:00 采集
6.0
How people use Copilot for Health
2026年04月20日
· 04/20 12:00 采集
5.0
MARCH: Multi-Agent Radiology Clinical Hierarchy for CT Report Generation
2026年04月20日
· 04/20 12:00 采集
5.0
DeepER-Med: Advancing Deep Evidence-Based Research in Medicine Through Agentic AI
2026年04月20日
· 04/20 12:00 采集

Posts

MemEvoBench: Benchmarking Memory MisEvolution in LLM Agents

AI Agents and Hard Choices

GTA-2: Benchmarking General Tool Agents from Atomic Tool-Use to Open-Ended Workflows

LLMs Corrupt Your Documents When You Delegate

Chain-of-Thought Degrades Visual Spatial Reasoning Capabilities of Multimodal LLMs

Symbolic Guardrails for Domain-Specific Agents: Stronger Safety and Security Guarantees Without Sacrificing Utility

The Semi-Executable Stack: Agentic Software Engineering and the Expanding Scope of SE

HarmfulSkillBench: How Do Harmful Skills Weaponize Your Agents?

Subliminal Transfer of Unsafe Behaviors in AI Agent Distillation

LLM Reasoning Is Latent, Not the Chain of Thought

MEDLEY-BENCH: Scale Buys Evaluation but Not Control in AI Metacognition

LinuxArena: A Control Setting for AI Agents in Live Production Software Environments

LACE: Lattice Attention for Cross-thread Exploration

The World Leaks the Future: Harness Evolution for Future Prediction Agents

Bilevel Optimization of Agent Skills via Monte Carlo Tree Search

ReactBench: A Benchmark for Topological Reasoning in MLLMs on Chemical Reaction Diagrams

Integrating Graphs, Large Language Models, and Agents: Reasoning and Retrieval

How people use Copilot for Health

MARCH: Multi-Agent Radiology Clinical Hierarchy for CT Report Generation

DeepER-Med: Advancing Deep Evidence-Based Research in Medicine Through Agentic AI