Topic: model spec

OpenAI Tackles Bias in ChatGPT

OpenAI is actively reducing political bias in ChatGPT, with its latest GPT-5 models showing significant improvements in objectivity and a 30% lower bias score compared to older versions. The company tested bias using a framework that evaluated responses to politically charged prompts from liberal...