Bias in Text Embedding Models