Maze Solving on Drunkard maze (test)
61.5Plan AccuracySolution-only
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Solution-onlyBackbone=Qwen3-8B-base, Fine-tuning regime=Solution-only2025.05 | 61.5 | 0 | |
| Correct TracesBackbone=Qwen3-8B-base, Fine-tuning regime=Correct Traces2025.05 | 53.4 | 33.7 | |
| Swapped TracesBackbone=Qwen3-8B-base, Fine-tuning regime=Swapped Traces2025.05 | 38.1 | 0 | |
| Base ModelBackbone=Qwen3-8B-base, Fine-tuning regime=None2025.05 | 14.5 | 0 |