failed_agent_trajectory

Mobile Trajectory Verification Data This dataset is prepared for verifying how LoT agent can help us identify incorrect actions taken in mobile execution trajectory. As mentioned in our paper (see citation 1), We select 52 mobile agent trajectories with specific task goals from instruction- guided executions on MagicWand platform(see citation 2). Each folder has a execution trajectory that includes screenshots along this trajectories, json files describing the details of each… See the full description on the dataset page: https://huggingface.co/datasets/orlando23/failed_agent_trajectory.