Sequential Crafting on TextCraft-3
82.5Success Rate (%)ReAct
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| ReActn-shot summary setting=none, LLM Backbone=gpt-4o-mini2026.04 | 82.5 | 16 | 5.2 | |
| Gen-Cycn-shot summary setting=none, LLM Backbone=gpt-4o-mini2026.04 | 71.5 | 22.1 | 3.2 | |
| Spec-Cycn-shot summary setting=none, LLM Backbone=gpt-4o-mini2026.04 | 50.4 | 21.6 | 2.3 | |
| DepDAGn-shot summary setting=none, LLM Backbone=gpt-4o-mini2026.04 | 17.9 | 15.4 | 1.2 |