Add location of unscored location/date file to README

README.md

@@ -38,7 +38,7 @@ The JSON file will live in the `auxiliary-data/modeled-clades/` directory of the
 This clade selection is based on the ["full open" NextStrain sequence metadata files](https://docs.nextstrain.org/projects/ncov/en/latest/reference/remote_inputs.html#remote-inputs-open-files), in particular [this file](https://data.nextstrain.org/files/ncov/open/metadata.tsv.zst) which is loaded and analyzed using [this script](https://github.com/reichlab/virus-clade-utils/blob/main/src/virus_clade_utils/get_clade_list.py). The NextStrain files are [typically updated daily in the late evening US eastern time](https://github.com/nextstrain/forecasts-ncov/actions/workflows/update-ncov-open-clade-counts.yaml) (it is only updated when new data are available). The hub pulls the most recent version of the file when the workflow runs each week. The precise lineage assignment model (sometimes referred to as a “reference tree”) that was used as well as the version of raw sequence data is stored as metadata, to facilitate reproducibility and evaluation. 
 
 ### Tasks for primary evaluation
-As described [below](#eval-challenges), only certain model tasks will be included in the primary model evaluation. These will include all clade frequencies for location-date pairs for which there are no observed specimens reported as of Wednesday night. A file that specifies which location-date pairs will be eligible for inclusion in the primary analysis will be generated and stored in the hub repository after the submission deadline passes.
+As described [below](#eval-challenges), only certain model tasks will be included in the primary model evaluation. These will include all clade frequencies for location-date pairs for which there are no observed specimens reported as of Wednesday night. A file that specifies which location-date pairs will be eligible for inclusion in the primary analysis will be generated and stored in the hub's `auxiliary-data/unscored-location-dates` directory after the submission deadline passes.
 
 ### Target data for evaluation
 Ninety days after each round closes, a script will generate a file containing summarized counts of selected clades for that round (including "other") for each location and date in the prediction window. These clade assignments will be made using the reference tree that was current when the submission round was open three months prior. While such "target data" files will not be suitable for training models (they will contain only limited dates and aggregated clades), they will be used as snapshots for evaluation.