1 paper
Zhan Su, Fengran Mo, Jinghan Zhang +4
The \textit{de facto} paradigm for applying dense retrieval (DR) to new tasks involves fine-tuning a pre-trained model for a specific task. However, this paradigm has two significa…