Text-based Task Completion on Jericho
3.37Mean Normalised ScoreReAct
Evaluation Results
| Method | Links | |
|---|---|---|
| ReActModel=Mistral Small 22B2026.04 | 3.37 | |
| DORAModel=Mistral Small 22B2026.04 | 2.92 | |
| ReActModel=Llama-3.1 8B2026.04 | 2.88 | |
| Chain of ThoughtModel=Llama-3.1 8B2026.04 | 2.22 | |
| Zero-shotModel=Llama-3.1 8B2026.04 | 2.21 | |
| DORAModel=Llama-3.1 8B2026.04 | 2.17 | |
| Tree of ThoughtModel=Llama-3.1 8B2026.04 | 2.05 | |
| ReActModel=Qwen-2.5 7B2026.04 | 1.87 | |
| Prompt ExploreModel=Mistral Small 22B2026.04 | 1.7 | |
| Prompt ExploreModel=Llama-3.1 8B2026.04 | 1.6 | |
| Chain of ThoughtModel=Qwen-2.5 7B2026.04 | 1.58 | |
| Chain of ThoughtModel=Mistral Small 22B2026.04 | 1.43 | |
| DORAModel=Qwen-2.5 7B2026.04 | 1.42 | |
| Tree of ThoughtModel=Qwen-2.5 7B2026.04 | 1.4 | |
| Zero-shotModel=Mistral Small 22B2026.04 | 1.35 | |
| Prompt ExploreModel=Qwen-2.5 7B2026.04 | 1.27 | |
| Tree of ThoughtModel=Mistral Small 22B2026.04 | 1.27 | |
| Zero-shotModel=Qwen-2.5 7B2026.04 | 0.66 |