Natural Language to SQL on Oncology clinical trial queries
94.64CHRFgpt-5m
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| gpt-5mStrategy=Zero-Shot, Backbone=gpt-5m2026.04 | 94.64 | 38.33 | 47.36 | 51.21 | — | — | |
| gpt-5mStrategy=Few-Shot, Backbone=gpt-5m2026.04 | 94.52 | 31.27 | 46.85 | 83.03 | — | — | |
| FD-NL2SQLStrategy=Ours, Backbone=gpt-5m2026.04 | 92.68 | 39.2 | 55.85 | 87.14 | — | — | |
| FD-NL2SQLStrategy=Ours, Backbone=Gemma32026.04 | 92.62 | 32.6 | 44.48 | 86.21 | 70.48 | 57.88 | |
| FD-NL2SQLStrategy=Ours, Backbone=Qwen32026.04 | 92.02 | 40.93 | 55.57 | 86.36 | 66.3 | 66.06 | |
| Gemma3Strategy=Few-Shot, Backbone=Gemma32026.04 | 91.32 | 30.85 | 45.72 | 85.47 | 53.16 | 56.8 | |
| FD-NL2SQLStrategy=Ours, Backbone=gpt-5n2026.04 | 90.89 | 32.53 | 50.25 | 86.32 | — | — | |
| gpt-5nStrategy=Few-Shot, Backbone=gpt-5n2026.04 | 89.58 | 15.8 | 31.97 | 75.53 | — | — | |
| gpt-5nStrategy=Zero-Shot, Backbone=gpt-5n2026.04 | 89.13 | 11.6 | 18.87 | 52.58 | — | — | |
| gpt-5nStrategy=CoT, Backbone=gpt-5n2026.04 | 86.93 | 13.67 | 33.47 | 77.56 | — | — | |
| gpt-5mStrategy=CoT, Backbone=gpt-5m2026.04 | 86.28 | 10.33 | 19.87 | 82.02 | — | — | |
| Gemma3Strategy=Zero-Shot, Backbone=Gemma32026.04 | 84.07 | 29.48 | 39.95 | 59.41 | 65.69 | 49.95 | |
| Qwen3Strategy=Few-Shot, Backbone=Qwen32026.04 | 83.02 | 25.6 | 52.95 | 80.43 | 34.7 | 53.15 | |
| Qwen3Strategy=Zero-Shot, Backbone=Qwen32026.04 | 81.82 | 26.9 | 46.51 | 72.54 | 39.01 | 51.58 | |
| Qwen3Strategy=CoT, Backbone=Qwen32026.04 | 77.96 | 21.65 | 29.73 | 85.21 | 0 | 0 | |
| Gemma3Strategy=CoT, Backbone=Gemma32026.04 | 77.89 | 13.47 | 34.88 | 0 | 0.25 | 0.18 | |
| SQL-R1Strategy=Zero-Shot, Backbone=SQL-R12026.04 | 63.78 | 3.32 | 1.75 | 41.78 | 10.39 | 0.42 | |
| SQL-R1Strategy=CoT, Backbone=SQL-R12026.04 | 50.24 | 2.85 | 3.35 | 49.15 | 1.69 | 0.15 |