Decision Making on Education Domain
78.8Top-1 AccuracyDECISIVE
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| DECISIVEDialogue=true2026.04 | 78.8 | 91.6 | 0.893 | 0.879 | 5.1 | |
| LLM-CoT (B2)Dialogue=false2026.04 | 65.5 | 83.5 | 0.804 | 0.786 | — | |
| LLM-Direct (B1)Dialogue=false2026.04 | 64.8 | 82.4 | 0.799 | 0.782 | — | |
| DeLLMaDialogue=false2026.04 | 28.1 | 44.3 | 0.733 | 0.48 | — | |
| DecisionFlowDialogue=false2026.04 | 20.2 | 34.3 | 0.706 | 0.403 | — | |
| LLM-Free Dialogue (B4)Dialogue=true2026.04 | 19.4 | 34.7 | 0.345 | 0.398 | 2.8 | |
| Active Pref. InferenceDialogue=true2026.04 | 13 | 25.9 | 0.679 | 0.336 | 9.7 | |
| LLM-Structured Dialogue (B3)Dialogue=true2026.04 | 9.4 | 22.2 | 0.225 | 0.302 | 4.7 |