What is Perplexity (Metric)?

Perplexity (Metric)

Definition A measurement of how well a language model predicts text — lower perplexity means the model is more confident and accurate in its predictions.

In Depth

Perplexity is calculated on a test dataset the model hasnt seen. A model with perplexity of 10 is essentially choosing between 10 equally likely next words on average. Lower is better. It is one of the most fundamental metrics for evaluating language models during training.

Browse more terms

AI Agent AI Alignment Agentic AI Artificial General Intelligence Artificial Superintelligence Attention Mechanism AutoML Autoregressive Model Benchmark Bias (AI) CLIP Chain-of-Thought Computer Vision Constitutional AI Context Window DPO (Direct Preference Optimization) Deepfake Diffusion Model Document AI Edge AI