Diversity Parameter

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

• Diversity Thresholdconcept0.816
  Target minimum diversity level (e.g., 10 contradictions) that DTG iterates toward
• Output Diversityconcept0.772
  The breadth of distinct outputs an LLM can produce, which is reduced by mode collapse after alignment training
• Task-specific parameterconcept0.765
  Parameter specific to each task, e.g., task head.
• Semantic Diversityconcept0.751
  Core concept measured by NLI Diversity — diversity of meaning across a set of dialogue responses
• Diversity Threshold Generationmethod0.750
  Iterative generation procedure that resamples lowest-scoring responses until a diversity threshold is reached
• Multi-Response Diversityconcept0.745
  Diversity measured over a set of m responses for a single conversation, as opposed to test set diversity
• Lexical Diversityconcept0.741
  Contrasted with semantic diversity; measured by n-gram overlap and distinct-n
• Test Set Diversityconcept0.739
  Traditional diversity measurement over one response per conversation across the test set