Skip to main content
. Author manuscript; available in PMC: 2024 Apr 30.
Published in final edited form as: Proc Conf Assoc Comput Linguist Meet. 2023 Jul;2023:10520–10542. doi: 10.18653/v1/2023.acl-long.587

Figure 1:

Figure 1:

Strategies for selecting rank sets of size 4 from larger candidate pools for relevance calibration (top half). The bottom half shows similar strategies to form binary contrast sets (2 positive, 2 negative) for faithfulness. Each strategy for the top half of the Figure occupies a row in Table 6, while the bottom corresponds to rows in Table 7.