Thanks for your work on releasing and maintaining the TimeQA dataset. I’m using the provided TimeQA dataset and noticed what seems to be a question–evidence mismatch in some samples.
Example
IDs: 3047–3077 (range where the issue is repeatedly observed)
Question topic: asks about Jörg Widmann’s position/tenure (employment/appointment information)
Problem: the supporting chunk/evidence associated with the answer appears to be about Anthony Fauci, not Jörg Widmann.
Thanks for your work on releasing and maintaining the TimeQA dataset. I’m using the provided TimeQA dataset and noticed what seems to be a question–evidence mismatch in some samples.
Example
IDs: 3047–3077 (range where the issue is repeatedly observed)
Question topic: asks about Jörg Widmann’s position/tenure (employment/appointment information)
Problem: the supporting chunk/evidence associated with the answer appears to be about Anthony Fauci, not Jörg Widmann.