measurement
The framework proposed in 'Grounding LLM Reasoning with Knowledge Graphs' achieved state-of-the-art performance on GRBench, a benchmark for domain-specific graph reasoning, with at least a 26.5% improvement over Chain-of-Thought (CoT) baselines.

Authors

Sources

Referenced by nodes (2)