Default LLM version for replication

[IsabelleQin]

Hi,

We are trying to replicate the results in your paper: Demonstration-Free: Towards More Practical Log Parsing with Large Language Models. As documented in the paper, the default LLM used is GPT-3.5-Turbo-0125. However, the default model used in the code is GPT-4o-mini. Can you please indicate which model we should use to replicate the results?

Thanks and have a nice day:-)

[yixiao2001]

sorry for the late reply, to run LogBatcher on large-scale datasets, we update the code and adopt GPT-4o-mini as the default model since it was the latest model and relatively cheap yet powerful. You can set either model as default LLM.