measurement
Chain-of-Thought (CoT) reasoning demonstrated significant improvements in hallucination mitigation in 71% of the models tested (p < 0.05), with 64% retaining significance after Benjamini-Hochberg FDR correction (q < 0.05).

Authors

Sources

Referenced by nodes (1)