Proactive Dialogue on ESConv
73.85Success RatePRINCIPLES
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| PRINCIPLESStrategy Set Size (|S|)=~1002025.09 | 73.85 | 6.36 | |
| AnEStrategy Set Size (|S|)=∞2025.09 | 58.46 | 7.38 | |
| ICL-AIFStrategy Set Size (|S|)=∞2025.09 | 56.15 | 7.87 | |
| Standard + GPT-4oStrategy Set Size (|S|)=02025.09 | 55.83 | 8.13 | |
| PPDPPStrategy Set Size (|S|)=8-162025.09 | 50.77 | 8.16 | |
| Standard + GPT-3.5-TurboStrategy Set Size (|S|)=02025.09 | 41.54 | 8.59 | |
| Proactive + MI-PromptStrategy Set Size (|S|)=8-162025.09 | 37.69 | 8.93 | |
| ProCoT + MI-PromptStrategy Set Size (|S|)=8-162025.09 | 35.38 | 9.13 | |
| ProactiveStrategy Set Size (|S|)=8-162025.09 | 23.85 | 9.51 | |
| ProCoTStrategy Set Size (|S|)=8-162025.09 | 22.31 | 9.51 | |
| NRPA-GD level 2Backbone=GPT-4o mini2026.04 | 1 | 4.17 | |
| UP-NRPA level 1Backbone=GPT-4o mini2026.04 | 1 | 3.81 | |
| UP-NRPA level 2Backbone=GPT-4o mini2026.04 | 1 | 3.25 | |
| UP-NRPA level 1Backbone=Qwen2.5 14B2026.04 | 1 | 2.92 | |
| UP-NRPA level 2Backbone=Qwen2.5 14B2026.04 | 1 | 2.76 | |
| DPDP -System 2Backbone=GPT-3.5-Turbo2026.04 | 0.9923 | 2.13 | |
| DPDP -System 1 & 2Backbone=GPT-3.5-Turbo2026.04 | 0.9923 | 2.13 | |
| DialogXpertBackbone=Qwen2.5 14B2026.04 | 0.9876 | 2.31 | |
| DialogXpertBackbone=Qwen 1.8B2026.04 | 0.9805 | 2.49 | |
| DialogXpertBackbone=Vicuna 13B2026.04 | 0.9651 | 2.7 | |
| DialogXpert -w/o EmotionBackbone=Qwen2.5 14B2026.04 | 0.9611 | 3.08 | |
| NRPA-GD level 1Backbone=GPT-4o mini2026.04 | 0.9461 | 5.28 | |
| DialogXpert -w/o RLBackbone=Qwen2.5 14B2026.04 | 0.9401 | 3.53 | |
| DialogXpert -w/o LLM-PriorBackbone=Qwen2.5 14B2026.04 | 0.9243 | 3.89 | |
| DPDP (System 1)Backbone=GPT-3.5-Turbo2026.04 | 0.9 | 3.61 | |
| DialogXpert -w/o RLBackbone=Qwen 1.8B2026.04 | 0.8921 | 4.04 | |
| DPDP -System 1 w/o PTBackbone=GPT-3.5-Turbo2026.04 | 0.8769 | 4.22 | |
| DialogXpert -w/o RLBackbone=Vicuna 13B2026.04 | 0.8749 | 4.13 | |
| DPDP -System 1 w/o SPTBackbone=GPT-3.5-Turbo2026.04 | 0.8692 | 3.97 | |
| PPDPPBackbone=Vicuna 13B2026.04 | 0.8462 | 4.56 | |
| PPDPP -w/o SFTBackbone=Vicuna 13B2026.04 | 0.8384 | 4.68 | |
| UDPBackbone=GPT-4o mini2026.04 | 0.832 | 7.59 | |
| ICL-AIFBackbone=-2026.04 | 0.8079 | 4.69 | |
| AnEBackbone=-2026.04 | 0.8 | 4.76 | |
| ProCoTBackbone=-2026.04 | 0.7923 | 4.75 | |
| Proactive + MI-PromptBackbone=-2026.04 | 0.7846 | 4.78 | |
| ProCoT + MI-PromptBackbone=-2026.04 | 0.7769 | 4.83 | |
| UDP -w/o PTBackbone=GPT-4o mini2026.04 | 0.772 | 7.48 | |
| StandardBackbone=-2026.04 | 0.7692 | 5.1 | |
| DialoGPTBackbone=GPT-22026.04 | 0.7538 | 5.31 | |
| ProactiveBackbone=-2026.04 | 0.7538 | 5.08 | |
| PPDPP -w/o RLBackbone=Vicuna 13B2026.04 | 0.7308 | 5.24 | |
| UDP -w/o RLBackbone=GPT-4o mini2026.04 | 0.531 | 8.64 |