The interpretability of the ReLU network to solve the problem of political correctness in the Black Myth of Wukong
Keywords: relu nlp wukong
Abstract: This article addresses the influence of political correctness on the Chinese game Black Myth: Wukong, particularly through the lens of feminist criticism. Using Natural Language Processing (NLP) enhanced with ReLU networks, we explore how interpretability can be improved to analyze gender bias objectively within media content. Traditional NLP models often rely on complex, opaque neural networks, which lack transparency in analyzing sensitive topics like feminism. By transforming deep networks into more interpretable shallow networks with ReLU, this study seeks to provide insights into bias detection and offer a framework for evaluating feminist discourse. Through this model, we aim to illuminate how bias within game narratives can be detected and addressed, fostering a more inclusive gaming culture. The paper concludes by discussing potential future developments in NLP interpretability and ethical AI in gender bias analysis.
Submission Number: 27
Loading