Thinking Fair and Slow: On the Efficacy of Structured Prompts for Debiasing Language Models

• https://arxiv.org/abs/2405.10431
• Shaz Furniturewala, Surgan Jandial, Abhinav Java, Pragyan Banerjee, Simra Shahid, Sumit Bhatia, Kokil Jaidka

LLMs, Bias

We evaluate a comprehensive end-user-focused iterative framework of debiasing that applies System 2 thinking processes … the more complex System 2-based Implicative Prompts significantly improve over other techniques demonstrating lower mean bias in the outputs with competitive performance on the downstream tasks.