List your company

Executable World-Model Learning on Baba in Wonderland extra-level transitions (held-out)

95.6Overall AccuracyAlice
69.39276.1968389.804May 16, 2026

Evaluation results

MethodLinks
AliceEvaluation Protocol=Offline, Calls=100, LLM Backbone=GPT-5.42026.0595.693.2
CWMEvaluation Protocol=Offline, Calls=100, LLM Backbone=GPT-5.42026.0582.673.4
GIF-MCTSEvaluation Protocol=Offline, Calls=100, LLM Backbone=GPT-5.42026.0570.462.8