philipphager/baidu-ultr_tencent-mlm-ctr

Query-document vectors and clicks for a subset of the Baidu Unbiased Learning to Rank dataset: https://arxiv.org/abs/2207.03051 This dataset uses the pretrained BERT cross-encoder (Bert_Layer12_Head12) from Tencent published as part of the WSDM cup 2023 to compute query-document vectors (768 dims): https://github.com/lixsh6/Tencent_wsdm_cup2023/tree/main/pytorch_unbias We link the model checkpoint also under `model/`.