which 16 models to generate the human annotation results

[dzf0023]

Hello authors,
Thank you so much for bringing these important resources to the community.

I have a question regarding the 16 models you used to generate the human annotations. In the paper you mentioned there were 16 models, I believe they its 4 extractive + 12 abstractive methods(section 4.3). But I did not see which 12 abstractive methods you used. In Table 3, there are 13 abstractive model results, which make me a little confused.

Could you please clarity which 12 abstractive models you use to get the human annotations? Thank you so much!