https://terms-en.ai-term-hub.com/en/tags/phenomena/Recent content in Phenomena on English AI Terms DictionaryHugoen-usSat, 18 Jul 2026 11:44:44 +0000https://terms-en.ai-term-hub.com/en/terms/grokking/Sat, 18 Jul 2026 10:00:30 +0000https://terms-en.ai-term-hub.com/en/terms/grokking/<h2 id="definition">Definition</h2> <p>Grokking refers to a counter-intuitive behavior observed in deep learning where a model continues to overfit on training data for a long time, showing poor generalization, before suddenly achieving near-perfect accuracy on both training and test sets. This delayed generalization typically occurs after thousands of epochs, suggesting that the network initially memorizes the data before discovering underlying patterns. It highlights the complex dynamics of optimization landscapes and the relationship between memorization and generalization in neural networks.</p>