Pre-training with general-reason and activity-specific facts improves task functionality devoid of hurting other model abilitiesAdditionally they enable The mixing of sensor inputs and linguistic cues within an embodied framework, improving decision-earning in true-planet scenarios. It boosts the model’s overall performance throughout many embodi