Can LLMs Effectively Leverage Graph Structural Information through Prompts, and Why?

Open paper
Benchmark Model Rank Results
text-to-sql-on-birdGPT-4 (Baseline)#6Execution Accuracy % (Test): 54.89Execution Accuracy % (Dev): 46.35
text-to-sql-on-birdClaude-2 (Baseline)#7Execution Accuracy % (Test): 49.02Execution Accuracy % (Dev): 42.70

Rank counts only results with a code link.