1 paper
Yongchang Peng, Qingshui Gu, Liya Zhu +31
Large language models (LLMs) are increasingly used for everyday assistance, yet existing benchmarks only partially reflect the requests users naturally make in practice. Real-world…