Fig. 2
From: Hard label adversarial attack with high query efficiency against NLP models
Comparison of query counts for texts with different lengths. The horizontal axis represents the collections of texts with different lengths, and the vertical axis represents the average query counts for each group.
