{"count":15,"terms":[{"slug":"embedding","term":"Embedding","short":"A list of numbers that represents the meaning of a piece of data, so that similar things end up with similar numbers.","category":"llms","status":"draft","url":"https://aidictionary.chanekmongkol.com/terms/embedding"},{"slug":"fine-tuning","term":"Fine-tuning","short":"Training an already-trained model a little more on a smaller, focused dataset so it gets better at a specific task or style.","category":"training","status":"draft","url":"https://aidictionary.chanekmongkol.com/terms/fine-tuning"},{"slug":"hallucination","term":"Hallucination","short":"When an AI model produces information that sounds confident and plausible but is false or made up.","category":"llms","status":"draft","url":"https://aidictionary.chanekmongkol.com/terms/hallucination"},{"slug":"large-language-model","term":"Large language model","short":"A very large neural network trained on huge amounts of text to predict the next piece of text, which lets it write, summarize, and answer questions.","category":"llms","status":"draft","url":"https://aidictionary.chanekmongkol.com/terms/large-language-model"},{"slug":"machine-learning","term":"Machine learning","short":"A way of building software that learns patterns from examples instead of following rules written by hand.","category":"basics","status":"draft","url":"https://aidictionary.chanekmongkol.com/terms/machine-learning"},{"slug":"neural-network","term":"Neural network","short":"A model made of layers of simple connected units that pass numbers forward, loosely inspired by neurons in the brain.","category":"basics","status":"draft","url":"https://aidictionary.chanekmongkol.com/terms/neural-network"},{"slug":"overfitting","term":"Overfitting","short":"When a model memorizes its training data instead of learning the general pattern, so it scores well in training but poorly on new data.","category":"training","status":"draft","url":"https://aidictionary.chanekmongkol.com/terms/overfitting"},{"slug":"oversampling","term":"Oversampling","short":"Repeating or creating more examples of a rare class so a dataset is more balanced before training.","category":"data","status":"draft","url":"https://aidictionary.chanekmongkol.com/terms/oversampling"},{"slug":"prompt-injection","term":"Prompt injection","short":"An attack where someone hides instructions in text a language model reads, so the model follows the attacker's instructions instead of the user's.","category":"safety","status":"draft","url":"https://aidictionary.chanekmongkol.com/terms/prompt-injection"},{"slug":"regularization","term":"Regularization","short":"Techniques that discourage a model from becoming too complex during training, which helps it avoid overfitting.","category":"training","status":"draft","url":"https://aidictionary.chanekmongkol.com/terms/regularization"},{"slug":"rag","term":"Retrieval-augmented generation","short":"A technique where a language model first looks up relevant documents, then writes its answer using them, so answers can be grounded in real sources.","category":"llms","status":"draft","url":"https://aidictionary.chanekmongkol.com/terms/rag"},{"slug":"token","term":"Token","short":"The small chunk of text a language model reads and writes, often a word, part of a word, or a punctuation mark.","category":"llms","status":"draft","url":"https://aidictionary.chanekmongkol.com/terms/token"},{"slug":"transformer","term":"Transformer","short":"A neural network design that uses attention to weigh how every part of the input relates to every other part. Most modern language models are built on it.","category":"llms","status":"draft","url":"https://aidictionary.chanekmongkol.com/terms/transformer"},{"slug":"underfitting","term":"Underfitting","short":"When a model is too simple or trained too little to capture the pattern in the data, so it does poorly even on its training data.","category":"training","status":"draft","url":"https://aidictionary.chanekmongkol.com/terms/underfitting"},{"slug":"validation-set","term":"Validation set","short":"A slice of data kept out of training and used to check the model and tune its settings while you build it.","category":"data","status":"draft","url":"https://aidictionary.chanekmongkol.com/terms/validation-set"}]}