Hi,
I am confused by phase 1 and phase 2 annotations in the dataset files.
The paper says phase 1 is a paraphrasing phase while phase 2 is a verification phase. As far as I understand, phase 2 is just about discarding wrong paraphrases. So what do you mean by a given example was collected in phase 1 vs. phase 2?
Thanks,
This is rather unfortunate naming but the phase number in the actual dataset indicates which batch the data was collected in. It is different than the phase in the paper.
Hi, I am confused by phase 1 and phase 2 annotations in the dataset files. The paper says phase 1 is a paraphrasing phase while phase 2 is a verification phase. As far as I understand, phase 2 is just about discarding wrong paraphrases. So what do you mean by a given example was collected in phase 1 vs. phase 2? Thanks,
--Ahmed