1 paper
Rongwu Xu, Zi'an Zhou, Tianwei Zhang +5
The common toxicity and societal bias in contents generated by large language models (LLMs) necessitate strategies to reduce harm. Present solutions often demand white-box access t…