Short definition
The tendency of large language models to reflect or amplify societal or dataset-based biases.
Definition
Definition
The tendency of large language models to reflect or amplify societal or dataset-based biases. An LLM that associates leadership roles with men more than women demonstrates bias.