Persuasive Dialogue on ToM-BPD interactive evaluation
55.23Win Rate: IdentificationQwen3-8B + TTBYS vs. GPT-5
Evaluation results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Qwen3-8B + TTBYS vs. GPT-5Base Model=Qwen3-8B, System=TTBYS, Comparison=vs. GPT-52026.05 | 55.23 | 26.47 | 34.56 | 25.12 | 42.78 | 26.31 | 35.44 | 29.18 | 28.62 | 31.04 | |
| Qwen3-8B + TTBYS vs. GPT-5 + CoTBase Model=Qwen3-8B, System=TTBYS, Comparison=vs. GPT-5 + CoT2026.05 | 45.19 | 29.44 | 37.33 | 25.78 | 42.51 | 36.29 | 33.22 | 27.65 | 30.11 | 28.47 | |
| GPT-5 vs. GPT-5 + CoTComparison=Head-to-head comparison2026.05 | 34.87 | 50.12 | 30.45 | 28.33 | 32.21 | 38.76 | 32.58 | 28.14 | 29.04 | 30.22 |