Text-based Task Completion on TextWorld
74.28Mean Normalised ScoreDORA
Evaluation results
| Method | Links | |
|---|---|---|
| DORAModel=Mistral Small 22B2026.04 | 74.28 | |
| Zero-shotModel=Mistral Small 22B2026.04 | 62.25 | |
| ReActModel=Mistral Small 22B2026.04 | 57.78 | |
| Prompt ExploreModel=Mistral Small 22B2026.04 | 50.6 | |
| DORAModel=Llama-3.1 8B2026.04 | 50.15 | |
| DORAModel=Qwen-2.5 7B2026.04 | 45.4 | |
| Prompt ExploreModel=Llama-3.1 8B2026.04 | 43.07 | |
| Zero-shotModel=Llama-3.1 8B2026.04 | 40.74 | |
| Chain of ThoughtModel=Llama-3.1 8B2026.04 | 37.72 | |
| Tree of ThoughtModel=Mistral Small 22B2026.04 | 32.54 | |
| ReActModel=Qwen-2.5 7B2026.04 | 31.43 | |
| Chain of ThoughtModel=Mistral Small 22B2026.04 | 31.43 | |
| Zero-shotModel=Qwen-2.5 7B2026.04 | 29.2 | |
| Chain of ThoughtModel=Qwen-2.5 7B2026.04 | 24.89 | |
| Prompt ExploreModel=Qwen-2.5 7B2026.04 | 23.12 | |
| ReActModel=Llama-3.1 8B2026.04 | 22.92 | |
| Tree of ThoughtModel=Llama-3.1 8B2026.04 | 15.06 | |
| Tree of ThoughtModel=Qwen-2.5 7B2026.04 | 11.39 |