AI
The Blind Curator: How a Biased Judge Silently Disables Skill Retirement in Self-Evolving Agents
A self-evolving agent retires its bad skills by watching them fail. However, if the judge cannot see...
AI
FormalAnalyticGeo: A Neural-Symbolic Based Framework for Multimodal Analytic Geometry Problem Generation
Researchers have developed a neural-symbolic framework called FormalAnalyticGeo to generate multimod...
AI
JUMP: Single-Pass Membership Inference on Fine-Tuned Diffusion Language Models
Researchers have proposed a new method called JUMP (Joint Uncertainty-Guided Mask Probing) to detect...
AI
OmegaUse-OfficeVal: Benchmarking LLM Agents on Long-Horizon Office-Suite Tasks with Economic Grounding
Researchers have created a benchmark called OmegaUse-OfficeVal to evaluate large language model (LLM...
AI
G-ReAct: Graph-Guided Deep Search via Structure-State Co-Evolution
Researchers have proposed a new framework for deep search called G-ReAct. It org...
AI
Agents Catching Agents: Shortcut Cascades and Benchmark Gaming in Clinical Multi-Agent Systems
A team of researchers has investigated the vulnerability of clinical decision su...
AI
VDGR-RAG: Vectors, Directories, Graphs, and Reflection Are All You Need for Unified Reasoning over Hierarchical Enterprise Knowledge
Researchers have developed VDGR-RAG, a unified framework for enterprise knowledg...
AI
Decided Upstream, Written Late: Locating and Pricing the Cross-Lingual Refusal Circuit of a Multilingual MoE
Researchers have identified a specific circuit in a multilingual model that caus...