The dataset identifies the shortcomings of existing benchmarks in evaluating the problem of compositional generalization, which underscores the need for the development of datasets tailored to assess compositional generalization in open intent detection tasks.
1 PAPER • 1 BENCHMARK