RITE: Evaluating End-to-End Resilience to Silent Tool Errors in E-Commerce Agentic Systems
This work forms Silent-error Resilience Task, an interactive task that evaluates whether agents complete the original query after a silent error corrupts an intermediate state, and introduces RITE, a benchmark built from human-authored industrial workflows and production-log-derived inputs.
Wen-Shuo Zhang, Yu-Quan Lin, Yu-Shi Sun et al.
· 0 citations