Checkr ditches GPT-4 for a smaller genAI model, streamlines background checks

In the pursuit of better accuracy and efficiency in background checks, Checkr's machine learning team realized that a smaller, fine-tuned language model could outperform larger models like GPT-4.

After struggling with GPT-4's 88% accuracy on background checks, Checkr pivoted to a small language model, fine-tuning it with years of accumulated data, achieving an impressive 97% accuracy rate.

Read at Computerworld