Decision Making on AlfWorld

6.4Steps

Teacher (LLaMA-13B)

Updated 1mo ago

Evaluation Results

Method
Teacher (LLaMA-13B) 2025.05	6.4	-	-	-
Structured Agent Distillation 2025.05	6.4	-	-	-
Teacher (OPT-13B) 2025.05	6.5	-	-	-
Structured Agent Distillation 2025.05	6.5	-	-	-
Structured Agent Distillation 2025.05	6.7	-	-	-
Token-level 2025.05	6.7	-	-	-
Token-level 2025.05	6.8	-	-	-
SeqKD 2025.05	6.9	-	-	-
Structured Agent Distillation 2025.05	7	-	-	-
SeqKD 2025.05	7	-	-	-
KD 2025.05	7.1	-	-	-
Token-level 2025.05	7.2	-	-	-
KD 2025.05	7.2	-	-	-
SeqKD 2025.05	7.5	-	-	-
Token-level 2025.05	7.8	-	-	-
KD 2025.05	7.8	-	-	-
SeqKD 2025.05	8	-	-	-
KD 2025.05	8.3	-	-	-
AutoRefine 2026.01	12.8	-	-	98.4
ReAct 2026.01	14	-	-	91
Reflexion 2026.01	15.7	-	-	95.5
ReAct + Reflexion 2026.01	16.1	-	-	95.5
ChatGPT o1-mini 2024.12	-	1.5	8.2	92.3
ChatGPT 4o 2024.12	-	14.2	20.9	76.6
ChatGPT 3.5-turbo 2024.12	-	7.5	5.2	76.5