Skip to Main Content
Table 7: 

Human fluency, entity-relevance, and faithfulness scores of entity-controlled models with the reference entities as the input. Faithful. denotes the percentage of generated summaries that are faithful. The Krippendorf’s α inter-rater agreement for all columns are 0.68, 0.77, and 0.56.

MethodFluencyEntity-rel.Faithful.
SD2 4.83 3.63 70% 
D.GPT2 4.65 3.33 68% 
D.GPT2+CMDP 4.83 3.92 71% 
MethodFluencyEntity-rel.Faithful.
SD2 4.83 3.63 70% 
D.GPT2 4.65 3.33 68% 
D.GPT2+CMDP 4.83 3.92 71% 
Close Modal

or Create an Account

Close Modal
Close Modal