4 papers
When Does Restricting a Coding Agent to execute_code Help? A Regime Agent-Design Ablation
Hong Yang, Qi Yu, Travis Desell
Modern coding agents expose multiple tool surfaces -- IDE primitives, bash, and Model Context Protocol (MCP) code-execution -- and the field has shipped three contradictory claims…
Beyond the Class Subspace: Teacher-Guided Training for Reliable Out-of-Distribution Detection in Single-Domain Models
Hong Yang, Devroop Kar, Qi Yu +2
Out-of-distribution (OOD) detection methods perform well on multi-domain benchmarks, yet many practical systems are trained on single-domain data. We show that this regime induces…
Domain Feature Collapse: Implications for Out-of-Distribution Detection and Solutions
Hong Yang, Devroop Kar, Qi Yu +2
Why do state-of-the-art OOD detection methods exhibit catastrophic failure when models are trained on single-domain datasets? We provide the first theoretical explanation for this…
Can We Ignore Labels In Out of Distribution Detection?
Hong Yang, Qi Yu, Travis Desell
Out-of-distribution (OOD) detection methods have recently become more prominent, serving as a core element in safety-critical autonomous systems. One major purpose of OOD detection…