Leap-Of-Thought: Teaching Pre-Trained Models to Systematically Reason Over Implicit Knowledge

Similar preprints — Semantic Scholar

Cited by (1)

• Evaluating Language Model Character Traits

  Claude-instant-1.2 achieves 91.1% accuracy and 88.6% logical coherence on 696 valid Leap-of-Thought entailment tuples — highest among 15 tested models including GPT-4 (89.9% accuracy, 84.7% coherence)