2 papers
cs.SE2026
Learning When and How to Intervene: A Hindsight-Distilled Sentinel for Coding Agents
Jiangrui Zhao, Chenglong Li, Meng Zhang +1
Coding agents solve repository-level tasks through sequences of actions, where a single erroneous action can misdirect subsequent decisions and increase recovery costs. Existing ap…
cs.CL2026
Syntax vs. Semantics: How Transformers Learn Deep Dependencies
Jiangrui Zhao, Xiaoting Du
Large Language Models demonstrate remarkable syntactic fluency, yet the optimization dynamics governing their acquisition of deep semantic dependencies remain poorly understood. We…