Multi-turn Dialogue on TSEData
96.46AccuracyChatAD-Mistral-7B
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| ChatAD-Mistral-7BBackbone=Mistral-7B2026.01 | 96.46 | 96.32 | 3.9 | |
| ChatAD*Backbone=Qwen2.5-7B, TKTO=true2026.01 | 96.07 | 95.83 | 2.19 | |
| ChatAD-Qwen2.5-7BBackbone=Qwen2.5-7B2026.01 | 95.69 | 95.39 | 1.95 | |
| ChatAD-Llama3-8BBackbone=Llama3-8B2026.01 | 82.38 | 83.96 | 27.7 | |
| TimeMQA-Llama3-8BBackbone=Llama3-8B2026.01 | 75.77 | 77.38 | 33.74 | |
| TimeMQA-Qwen2.5-7BBackbone=Qwen2.5-7B2026.01 | 70.3 | 71.74 | 37.32 | |
| Mistral-7B-InstructBackbone=Mistral-7B2026.01 | 70.17 | 75.7 | 53.03 | |
| Llama3-8B-InstructBackbone=Llama3-8B2026.01 | 69.63 | 73.88 | 48.66 | |
| Qwen2.5-7B-InstructBackbone=Qwen2.5-7B2026.01 | 69.42 | 73.31 | 47.32 | |
| TimeMQA-Mistral-7BBackbone=Mistral-7B2026.01 | 66.11 | 67.11 | 40.47 | |
| Qwen3-8BBackbone=Qwen3-8B2026.01 | 60.86 | 70.66 | 69.98 | |
| DeepSeek-R1-Distill-Llama3-8BBackbone=Llama3-8B, Type=Distilled2026.01 | 60.18 | 68.12 | 66.25 | |
| DeepSeek-R1-Distill-Qwen2.5-7BBackbone=Qwen2.5-7B, Type=Distilled2026.01 | 50.99 | 65.87 | 92.77 |