List your company
ResearchFormerly WizWand

Texas Hold'em

0.3DQN ScoreAgent-Pro w/o Learning
-4.172-3.011-1.85-0.689Feb 27, 2024
Updated 8mo ago

Evaluation results

MethodLinks
Agent-Pro w/o LearningBased Model=Llama2-70B2024.020.33.3-1.2-1.8
ReflexionBased Model=GPT3.52024.02-0.12.5-0.9-1.5
Vanilla LLMBased Model=GPT3.52024.02-0.32.2-0.8-1.1
Aggressive LLMBased Model=GPT3.52024.02-0.43-0.5-2.1
Self-ConsistencyBased Model=GPT3.52024.02-0.51.9-0.8-0.6
Conservative LLMBased Model=Llama2-70B2024.02-0.53.4-0.8-2.1
Conservative LLMBased Model=GPT3.52024.02-0.72.9-0.9-1.3
ReActBased Model=GPT3.52024.02-0.71.7-0.7-0.3
Agent-Pro w/o LearningBased Model=GPT3.52024.02-0.71.8-1-0.1
Multi-AgentBased Model=Llama2-70B2024.02-0.73.5-1-1.8
Vanilla LLMBased Model=Llama2-70B2024.02-0.83.4-0.4-2.2
Self-ConsistencyBased Model=Llama2-70B2024.02-13.8-0.9-1.9
Multi-AgentBased Model=GPT3.52024.02-1.12.3-0.3-0.9
ReActBased Model=Llama2-70B2024.02-1.13.9-0.8-2
ReflexionBased Model=Llama2-70B2024.02-1.24.7-0.9-2.6
Agent-ProBased Model=Llama2-70B2024.02-1.23.1-0.5-1.4
Agent-ProBased Model=GPT3.52024.02-1.51.4-1.11.2
Conservative LLMBased Model=GPT42024.02-1.62.7-1.60.5
Aggressive LLMBased Model=Llama2-70B2024.02-1.67.6-1.2-4.8
Multi-AgentBased Model=GPT42024.02-1.81.9-1.21.1
Aggressive LLMBased Model=GPT42024.02-22.8-10.2
Vanilla LLMBased Model=GPT42024.02-2.21.7-0.91.4
ReActBased Model=GPT42024.02-2.41.3-1.12.2
ReflexionBased Model=GPT42024.02-2.62.1-0.71.2
Self-ConsistencyBased Model=GPT42024.02-2.82-0.71.5
Agent-Pro w/o LearningBased Model=GPT42024.02-31.5-1.22.7
Agent-ProBased Model=GPT42024.02-3.91.1-1.54.3
HumanBased Model=GPT3.52024.02-40.7-2.45.7
HumanBased Model=GPT42024.02-40.7-2.45.7
HumanBased Model=Llama2-70B2024.02-40.7-2.45.7