1
Predicting Performance of Symbolic and Prompt Programs with Examples
提出基于示例预测LLM符号与提示程序性能的新方法，为智能编程提供高效评估框架。
arXiv:2605.21515v1 Announce Type: cross Abstract: LLM prompting is widely used for naturally stated tasks, yet it is unreliable it may succeed on a fe…