List your company

Social Navigation on Social Navigation Simulation 10 Human

80Success Rate (SR)NaviWM
-3.218.44061.6Oct 27, 2025

Evaluation results

MethodLinks
NaviWMBackbone=GPT-52025.108014.998.2372185
NaviWMBackbone=GPT-4o2025.106020.1516.687173
NaviWMBackbone=GPT-o42025.106012.197.4583171
GPT-o4 (mini)Backbone=GPT-o4 (mini)2025.105011.636.6399149
GPT-5Backbone=GPT-52025.104011.456.587170
Abl-WM-DeepSeekBackbone=DeepSeek2025.104010.354.581533
DeepSeek-V3Backbone=DeepSeek-V32025.10309.63.682847
GPT-4oBackbone=GPT-4o2025.103017.74770115
Abl-WM-GPTBackbone=GPT2025.103010.283.582137
Abl-WM-LLaMABackbone=LLaMA2025.102012.935.552045
Abl-CoT-GPTBackbone=GPT, Reasoning=Zero-Shot CoT2025.10208.793.63059
LLaMA-8BBackbone=LLaMA-8B2025.10026.8420.51023
LLaMA-405BBackbone=LLaMA-405B2025.1007.193.252652
GPT-3.5Backbone=GPT-3.52025.100043.253482
GPT-4Backbone=GPT-42025.10015.2927.669139