HT on HT
367Generation ScoreRand
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Randthreshold (tau)=0.70, Base LLM=Qwen3-8B2026.05 | 367 | 381.23 | |
| EXP3.Pthreshold (tau)=0.70, Base LLM=Qwen3-8B2026.05 | 209 | 213.22 | |
| ShinkaEvolvethreshold (tau)=0.50, Base LLM=Qwen3-8B2026.05 | 201 | 189.28 | |
| UCBthreshold (tau)=0.70, Base LLM=Qwen3-8B2026.05 | 125 | 130.55 | |
| EXP3.Pthreshold (tau)=0.50, Base LLM=Qwen3-8B2026.05 | 102 | 108.67 | |
| Thompsonthreshold (tau)=0.70, Base LLM=Qwen3-8B2026.05 | 101 | 104.53 | |
| Randthreshold (tau)=0.50, Base LLM=Qwen3-8B2026.05 | 92 | 99.9 | |
| Thompsonthreshold (tau)=0.50, Base LLM=Qwen3-8B2026.05 | 74 | 79.45 | |
| UCBthreshold (tau)=0.50, Base LLM=Qwen3-8B2026.05 | 60 | 66.9 |