Text-to-SQL Ranking Correlation on BIRD
-0.11Kendall's τLLaMA3.3-70B
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| LLaMA3.3-70BRank subset=All ranks2026.04 | -0.11 | -0.73 | |
| LLaMA3.3-70BRank subset=Ranks ≥ 32026.04 | -0.14 | -0.77 | |
| LLaMA3.1-405BRank subset=Ranks ≥ 32026.04 | -0.24 | -1 | |
| GPT-4o-miniRank subset=All ranks2026.04 | -0.33 | -0.82 | |
| O3-miniRank subset=Ranks ≥ 32026.04 | -0.36 | -1 | |
| O3-miniRank subset=All ranks2026.04 | -0.38 | -0.88 | |
| LLaMA3.1-405BRank subset=All ranks2026.04 | -0.39 | -0.88 | |
| GPT-OSS-20Rank subset=All ranks2026.04 | -0.42 | -0.87 | |
| GPT-OSS-20Rank subset=Ranks ≥ 32026.04 | -0.43 | -0.84 | |
| GPT-OSS-120Rank subset=All ranks2026.04 | -0.47 | -0.85 | |
| GPT-OSS-120Rank subset=Ranks ≥ 32026.04 | -0.5 | -1 | |
| GPT-4o-miniRank subset=Ranks ≥ 32026.04 | -0.64 | -1 |