Table 10.
Run | Passages | TP | Unique | U_TP | TP/Passages | U_TP/Unique | MRR |
Run #1 | 372 | 51 | 361 | 51 | 13.71 | 14.13 | 1.0 |
Run #2 | 372 | 71 | 361 | 70 | 19.09 | 19.39 | 1.0 |
Column headings: Passages = the total number of passages evaluated; TP = the number of evaluated passages that were preselected by human curators; Unique = the number of unique passages evaluated; U_TP, the number of unique passages that were pre-selected; MRR = mean reciprocal rank of the correct passages. ISS, interaction sentence subtask.