Scaling Implicit Bias Analysis across Transformer-Based Language Models through Embedding Association Test and Prompt Engineering

In the evolving field of machine learning, deploying fair and transparent models remains a formidable challenge. This study builds on earlier research, demonstrating that neural architectures exhibit inherent biases by analyzing a broad spectrum of transformer-based language models from base to x-la...

ver descrição completa

Detalhes bibliográficos

Principais autores:	Ravi Varma Kumar Bevara, Nishith Reddy Mannuru, Sai Pranathi Karedla, Ting Xiao
Formato:	Artigo
Idioma:	Inglês
Publicado em:	MDPI AG 2024-04-01
coleção:	Applied Sciences
Assuntos:	AI model scaling Generative Pretrained Transformer (GPT) k-means Bidirectional Encoder Representations from Transformers (BERT) transformer
Acesso em linha:	https://www.mdpi.com/2076-3417/14/8/3483
Tags:	Adicionar Tag Sem tags, seja o primeiro a adicionar uma tag!