Update README.md

README.md

@@ -23,7 +23,7 @@ co2_eq_emissions:
 A flax-based MonoBERT cross encoder trained on the [Baidu-ULTR](https://arxiv.org/abs/2207.03051) dataset with a **listwise DLA objective on clicks**. Following [Ai et al.](https://arxiv.org/abs/1804.05938), the dual learning algorithm jointly infers item relevance (using a BERT model) and position bias (in our case, a single embedding parameter per rank), both by optimizing a **listwise softmax cross-entropy loss**. For more info, [read our paper](https://arxiv.org/abs/2404.02543) and [find the code for this model here](https://github.com/philipphager/baidu-bert-model).
 
 ## Test Results on Baidu-ULTR
-Ranking performance is measured in DCG, nDCG, and MRR on expert annotations (6,985 queries). Click prediction performance is measured in log-likelihood on one test partition of user clicks (49,495 queries).
+Ranking performance is measured in DCG, nDCG, and MRR on expert annotations (6,985 queries). Click prediction performance is measured in log-likelihood on one test partition of user clicks (≈297k queries).
 
 | Model                                                                                          | Log-likelihood | DCG@1 | DCG@3 | DCG@5 | DCG@10 | nDCG@10 | MRR@10 |
 |------------------------------------------------------------------------------------------------|----------------|-------|-------|-------|--------|---------|--------|