Relevance Classification on Chinese Conversational AI Financial Services (test)
95AccuracyLatent Model
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Latent Modelaggregation=Query-Adaptive Latent Ensemble2026.01 | 95 | 94 | 95 | |
| BGE@0.85threshold=0.852026.01 | 83 | 80 | 82 | |
| Majority Vote2026.01 | 80 | 78 | 75 | |
| Single LLM2026.01 | 78 | 75 | 72 |