How can large language models alleviate tedious network-related tasks? We provide evaluations to understand the bottlenecks and align the language models with domain expertise.
Evaluating advanced language models (Llama 3.2, GPT4o) performance on classifying NSP’s types based on their website data, with textual and visual input.
Leveraging preference optimization techniques (SimPO, CPO) to improve the alignment of model classification logic with human-labeled ground truths.