What is the role of neutral NLI predictions in measuring semantic diversity?

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

• Neutral NLI predictions may capture lexical rather than semantic diversityhypothesis0.917
  Hypothesis proposed to explain Neutral NLI Diversity's high performance on decTest but low on conTest
• Confidence NLI Diversity achieves state-of-the-art performance on measuring semantic diversityclaim0.852
  Main performance claim of the paper
• Neutral NLI Diversitymethod0.815
  Variant weighting neutral predictions equally to contradictions to test if neutrals capture lexical diversity
• What is the effect of more generalized NLI training data on NLI Diversity performance?question0.804
  Future work question arising from the finding that Combined model did not outperform MNLI model
• NLI Diversity is limited by the scope of the NLI model's training dataclaim0.801
  Limitation acknowledged in discussion section
• The contradiction relation is more useful than the neutral relation for measuring semantic diversityclaim0.798
  Key finding from ablation study on NLI prediction categories
• 95% confidence intervals overlap between Baseline NLI Diversity, Confidence NLI Diversity, SentBERT, and human judgments on conTestfinding0.772
  Indicates lack of statistically significant differences between top methods
• Confidence NLI Diversity achieves ρ=0.64 correlation with human diversity judgments on conTestfinding0.771
  Highest human correlation for semantic diversity metric