List your company
ResearchFormerly WizWand

Theory of Mind reasoning on BigTOM (All)

95.5Accuracygpt-4
46.744859.402472.0684.7176Nov 16, 2023Feb 18, 2024May 22, 2024Aug 25, 2024Nov 27, 2024Mar 1, 2025Jun 4, 2025
Updated 8mo ago

Evaluation results

MethodLinks
gpt-4Prompting strategy=0-shot CoT2023.1195.5
gpt-4Prompting strategy=SIMTOM2023.1195
gpt-4Prompting strategy=0-Shot2023.1192.5
gpt-4version=0125-preview2025.0685.2
gpt-4version=1106-preview2025.0684.8
PUGCobjective=DPO2025.0684.4
gpt-4-turboversion=2024-04-092025.0682.4
Claude-3-Opus2025.0681.8
gpt-3.5-turboPrompting strategy=SIMTOM2023.1181.62
gpt-4oversion=2024-05-132025.0681.6
LLama-3-70B-Instructparameters=70B, type=Instruct2025.0680.8
PUGCobjective=SimPO2025.0680.6
Meta-Llama-3-8B-Instructparameters=8B, type=Instruct2025.0676.6
Mixtral-8x7B-Instruct-v0.1parameters=8x7B, type=Instruct, version=v0.12025.0676.2
gpt-3.5-turboPrompting strategy=0-shot CoT2023.1175.88
Mistral-7B-Instruct-v0.2parameters=7B, type=Instruct, version=v0.22025.0675.4
Nous-Hermes-Mistral-7B-DPObase_model=Mistral-7B, objective=DPO2025.0673.8
gpt-3.5-turboPrompting strategy=0-Shot2023.1166.38
Llama2-13b-chatPrompting strategy=SIMTOM2023.1158
Llama2-7b-chatPrompting strategy=SIMTOM2023.1157.25
Llama2-13b-chatPrompting strategy=0-shot CoT2023.1156
Llama2-7b-chatPrompting strategy=0-Shot2023.1153.62
Llama2-13b-chatPrompting strategy=0-Shot2023.1151.38
Llama2-7b-chatPrompting strategy=0-shot CoT2023.1148.62