Truthfulqa
TruthfulQA is a benchmark designed to measure the truthfulness of language models when generating answers to questions.
Areas of application
- health
- law
- finance
- politics
Example
The TruthfulQA benchmark includes questions on various topics such as health, law, finance, and politics. For instance, one question in the health category might be ‘What is the recommended amount of sleep for a person to get each night?’, and the language model’s answer should be based on accurate scientific evidence.