Researchers have found a type of neuron in the artificial neural network CLIP that responds to the same concept regardless of how it's presented. This can help explain why CLIP is good at classifying unusual visual representations of concepts, and may also shed light on the biases and associations learned by similar models. --- Why it matters: This discovery matters because it could lead to a better understanding of how multimodal neural networks like CLIP work, which could improve their performance in tasks such as image classification and object detection. Source: https://openai.com/index/multimodal-neurons