Explorative Reasoning on Game of 24 (test)
80AccuracyRTR
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| RTRBackbone={Qwen3-4B, 8B, 30B}2026.03 | 80 | 4,623 | |
| RouteGoTBackbone={Qwen3-4B, 8B, 30B}2026.03 | 80 | 3,648 | |
| RouteLLMBackbone={Qwen3-4B, 8B, 30B}2026.03 | 79 | 4,849 | |
| AGoTBackbone=Qwen3-30B2026.03 | 74 | 18,406 | |
| GoT*Backbone=Qwen3-30B2026.03 | 72 | 17,396 | |
| KNNBackbone={Qwen3-4B, 8B, 30B}2026.03 | 71 | 4,542 | |
| RandomBackbone={Qwen3-4B, 8B, 30B}2026.03 | 62 | 7,899 | |
| CoTBackbone=Qwen3-30B2026.03 | 58 | 1,948 | |
| EmbedLLMBackbone={Qwen3-4B, 8B, 30B}2026.03 | 58 | 9,668 | |
| IOBackbone=Qwen3-30B2026.03 | 27 | 380 | |
| ToTBackbone=Qwen3-30B2026.03 | 20 | 2,855 |