List your company

Web/tool tasks on WebArena 245 (test)

70.4Success RateCVT-RL
62.28864.39466.568.606Jun 3, 2026

Evaluation results

MethodLinks
CVT-RLDescription=per-seed CVT-RL success, Split (train/dev/test)=1850/230/245, Individual Seeds=68.3/71.5/70.1/69.7/72.42026.0670.43.5
IMDescription=information-matched counterfactual-process RL, Split (train/dev/test)=1850/230/2452026.0666.9—
CMDescription=compute-matched non-causal RL, Split (train/dev/test)=1850/230/2452026.0662.6—