Query Refinement Evaluation on e-commerce dataset
4.67Realism ScoreHuman
Evaluation Results
| Method | Links | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| HumanRefinement Category=Baseline, Evaluator Model=gemma3-27b2026.05 | 4.67 | — | — | 0.53 | 0 | 0.45 | 0.78 | 0.88 | 0 | 88.9 | 0 | |
| PQRRefinement Category=Query Refinement, Evaluator Model=gemma3-27b2026.05 | 4.49 | 48.11 | 1 | 0.63 | 18.87 | 0.36 | 0.81 | 0.95 | -2.73 | 117.05 | 31.66 | |
| PAIRRefinement Category=Query Refinement, Evaluator Model=gemma3-27b2026.05 | 3.52 | 44.12 | 1.8 | 0.69 | 30.19 | 0.31 | 0.71 | 0.89 | -12.98 | 145.12 | 63.24 | |
| TAPRefinement Category=Query Refinement, Evaluator Model=gemma3-27b2026.05 | 3.13 | 42.08 | 1.7 | 0.63 | 18.87 | 0.36 | 0.7 | 0.83 | -11.98 | 122.54 | 37.84 | |
| ACERefinement Category=Prompt Refinement, Evaluator Model=gemma3-27b2026.05 | 2.67 | 44.92 | 1.5 | 0.49 | -7.55 | 0.36 | 0.75 | 0.89 | -7.57 | 125.62 | 41.3 | |
| GEPARefinement Category=Prompt Refinement, Evaluator Model=gemma3-27b2026.05 | 2.59 | 31.67 | 1.7 | 0.68 | 28.3 | 0.25 | 0.62 | 0.78 | -25.44 | 86.92 | -2.23 |