Phase 2 types of memes in test

According to the paper, we know the test sets in phase1 have 5 types of memes, including 40% multimodal hate, 10% unimodal hate, 20% benign text confounder, 20% benign image confounder, 10% random non-hateful. Is Phase 2 the same as Phase1? Do Phase2 contain benign text and image confounder?