Research findings on ethical biases in leading language models
This article will present our research findings on the ethical biases exposed by leading LLMs when making decisions without explicit values guidance, and how VALUES.md can address these issues.