Cohere For AI - Guest Speaker: Dr. Shaina Raza, Applied ML Scientist

In addressing the critical need for safety in Large Language Models (LLMs), it is crucial to ensure that the outputs are not only safe but also retain their contextual accuracy. Many existing LLMs are safe fine-tuned either with safety demonstrations, or rely only on adversarial testing. While able to get safe outputs, they often risk losing contextual meaning as they mitigate bias and toxicity.

In response, we present MBIAS, a LLM framework instruction fine-tuned on a custom dataset specifically designed for safety interventions. MBIAS aims to address the significant issues of bias and toxicity in LLMs generations that typically manifest as underrepresentation or negative portrayals across various demographics, including inappropriate linguistic mentions and biased content in social media. We experiment on MBIAS for safety interventions using various configurations, and demonstrate more than a 30% reduction in overall bias and toxicity while successfully retaining key information.

Additionally, a demographic analysis on an out of-distribution test set confirms the robustness of our approach, with reductions in bias and toxicity exceeding 90% across various demographics. The dataset and instruction fine-tuned MBIAS are made available to the research community.

Bio: Dr. Shaina Raza currently is as an Applied Machine Learning Scientist with a focus on Responsible AI at the Vector Institute of Artificial Intelligence, located in Toronto, Canada. She holds a Ph.D. in Computer Science, where her research centered around Recommender Systems and Fairness in Machine Learning. Her research portfolio is diverse, covering areas such as Recommender Systems, ML Fairness, Public Health Equity, and Ethical AI. Before her current position, Dr. Raza enriched her experience as a Postdoctoral Fellow at Public Health Ontario and the Dalla Lana School of Public Health. She held the esteemed position of CIHR HSIF award holder during this time. Over the years, Dr. Raza has made several contributions to her academic field. She continues her commitment to understanding and innovating at the crossroads of technology and societal welfare.