A survey on fairness in large language models

Y Li, M Du, R Song, X Wang, Y Wang - arXiv preprint arXiv:2308.10149, 2023 - arxiv.org
Large language models (LLMs) have shown powerful performance and development
prospect and are widely deployed in the real world. However, LLMs can capture social …

The life cycle of large language models in education: A framework for understanding sources of bias

J Lee, Y Hicke, R Yu, C Brooks… - British Journal of …, 2024 - Wiley Online Library
Large language models (LLMs) are increasingly adopted in educational contexts to provide
personalized support to students and teachers. The unprecedented capacity of LLM‐based …

Palm 2 technical report

R Anil, AM Dai, O Firat, M Johnson, D Lepikhin… - arXiv preprint arXiv …, 2023 - arxiv.org
We introduce PaLM 2, a new state-of-the-art language model that has better multilingual and
reasoning capabilities and is more compute-efficient than its predecessor PaLM. PaLM 2 is …

Holistic evaluation of language models

P Liang, R Bommasani, T Lee, D Tsipras… - arXiv preprint arXiv …, 2022 - arxiv.org
Language models (LMs) are becoming the foundation for almost all major language
technologies, but their capabilities, limitations, and risks are not well understood. We present …

From pretraining data to language models to downstream tasks: Tracking the trails of political biases leading to unfair NLP models

S Feng, CY Park, Y Liu, Y Tsvetkov - arXiv preprint arXiv:2305.08283, 2023 - arxiv.org
Language models (LMs) are pretrained on diverse data sources, including news, discussion
forums, books, and online encyclopedias. A significant portion of this data includes opinions …

Biases in large language models: origins, inventory, and discussion

R Navigli, S Conia, B Ross - ACM Journal of Data and Information …, 2023 - dl.acm.org
In this article, we introduce and discuss the pervasive issue of bias in the large language
models that are currently at the core of mainstream approaches to Natural Language …

On the opportunities and risks of foundation models

R Bommasani, DA Hudson, E Adeli, R Altman… - arXiv preprint arXiv …, 2021 - arxiv.org
AI is undergoing a paradigm shift with the rise of models (eg, BERT, DALL-E, GPT-3) that are
trained on broad data at scale and are adaptable to a wide range of downstream tasks. We …

Evaluating the social impact of generative ai systems in systems and society

I Solaiman, Z Talat, W Agnew, L Ahmad… - arXiv preprint arXiv …, 2023 - arxiv.org
Generative AI systems across modalities, ranging from text, image, audio, and video, have
broad social impacts, but there exists no official standard for means of evaluating those …

Causal inference in natural language processing: Estimation, prediction, interpretation and beyond

A Feder, KA Keith, E Manzoor, R Pryzant… - Transactions of the …, 2022 - direct.mit.edu
A fundamental goal of scientific research is to learn about causal relationships. However,
despite its critical role in the life and social sciences, causality has not had the same …

On second thought, let's not think step by step! Bias and toxicity in zero-shot reasoning

O Shaikh, H Zhang, W Held, M Bernstein… - arXiv preprint arXiv …, 2022 - arxiv.org
Generating a Chain of Thought (CoT) has been shown to consistently improve large
language model (LLM) performance on a wide range of NLP tasks. However, prior work has …