Question

[Kledki]

In the github project, is the "data" folder the dataset you used for the final experiment?

[JinheonBaek]

Yes, your understanding is correct.

We use all the papers and references to build the entity-centric knowledge store.

For research idea generation, we randomly sample 300 papers with the seed number of 42 from the "papers.jsonl" file.