Perplexity (Metric) vs Perplexity
People think Perplexity the score and Perplexity the app are the same thing. They only share a name. One is a number that says how well a language model predicts text, and lower is better. The other is an AI search app that answers questions with links to its sources.
Perplexity (Metric)
A score for how well a language model predicts text. Lower means less surprised.
Perplexity measures how well a language model predicts text it has not seen. At each step, the model gives a chance to the token that really comes next. Perplexity turns all those chances into one number. A perplexity of 10 means the model is as unsure as if it were picking from 10 equally likely words. Lower is better. It checks how well the model knows the language, not whether its answers are true.
Perplexity = e to the power of (the average of -log(chance of each real next token))Perplexity
An AI search app that answers with links to its sources.
Perplexity is an AI answer engine. You ask a question, and it searches the web for you. Then it writes one short answer, with numbered links to the pages it used. You can click a number to check the source. It started in 2022 in San Francisco. It also has a mode for longer research and its own web browser.
In one line
- Perplexity (Metric)
- A score for how well a language model predicts text. Lower means less surprised.
- Perplexity
- An AI search app that answers with links to its sources.