BASFuzz, a fuzz‑testing framework for LLM‑based NLP apps, achieved 90.335% effectiveness and cut evaluation time by 2,163.852 seconds versus prior baselines. Read more: https://getnews.me/basfuzz-introduces-automated-fuzz-testing-for-llm-based-nlp-software/ #basfuzz #fuzztesting #nlp

Mingxuan Xiao, Yan Xiao, Shunhui Ji, Jiahe Tu, Pengcheng Zhang: BASFuzz: Towards Robustness Evaluation of LLM-based NLP Software via Automated Fuzz Testing https://arxiv.org/abs/2509.17335 https://arxiv.org/pdf/2509.17335 https://arxiv.org/html/2509.17335