 A generalization probe has nothing to do with aliens has everything to do with finding out whether or not the training that the behavior is generalizing to a setting outside of the training setting. All right? So for example, taught my kid Otis, well I didn't do it, but somebody taught my kid Otis to swim in the summer in a pool. It was great. He did a really good job. He didn't swim very well, but anyway he's learning. Okay, give the kid a break. So learning, he's swimming, he's doing his thing, and then we took him to the lake and we threw him in the water and we haven't seen him since. Generalization failure. Not because he swam so fast, but because he sunk so quick. I got popped in my head that maybe the kid took off across the lake. That would not be a generalization failure. It is still a probe, right? So generalization success would be he swam and he took off across the lake. Generalization failure would be he got into the lake and went, oh no! And it sank, right? He didn't sink. It was fine. It was a joke. He's totally here. I dropped him off at school this morning. It's totally good. So generalization probes, testing to see if the training is working in different contexts, different environments. I think that's about it. I think you get it. He didn't drown. He's fine.