Browse Teach Curate

Perplexity as a Language Model Evaluation Metric

Perplexity measures how well a probabilistic language model predicts a text as the exponential of the average negative log-likelihood per token; lower values indicate the text is more fluent or expected under the model, and it is widely used to score text quality.

jeremy Video

Evaluating Language Models Using Perplexity in Natural Language Processing

Perplexity serves as a fundamental evaluation metric in Natural Language Processing that quantifies how well a probability distribution assigned by a language model approximates the true underlying d…