A problem about the splitting "&&" in eval_sequences DIVERSE run_ner.py

[lyconghk]

For the task GSM8K, there is no detailed information about the splitting token "&&" in the data file. Can you explain where this '&&' comes from? I have already checked the train.txt file and the dataset construction, it seems that there is no information about this '&&' at all.

Thanks.