English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Generative Pre-trained Transformer (GPT) models have exhibited excitingprogress in their capabilities, capturing the interest of practitioners and thepublic alike. Yet, while the literature on the trustworthiness of GPT modelsremains limited, practitioners have proposed employing capable GPT models forsensitive applications such as healthcare and finance -- where mistakes can becostly. To this end, this work proposes a comprehensive trustworthinessevaluation for large language models with a focus on GPT-4 and GPT-3.5,considering diverse perspectives -- including toxicity, stereotype bias,adversarial robustness, out-of-distribution robustness, robustness onadversarial demonstrations, privacy, machine ethics, and fairness. Based on ourevaluations, we discover previously unpublished vulnerabilities totrustworthiness threats. For instance, we find that GPT models can be easilymisled to generate toxic and biased outputs and leak private information inboth training data and conversation history. We also find that although GPT-4is usually more trustworthy than GPT-3.5 on standard benchmarks, GPT-4 is morevulnerable given jailbreaking system or user prompts, potentially because GPT-4follows (misleading) instructions more precisely. Our work illustrates acomprehensive trustworthiness evaluation of GPT models and sheds light on thetrustworthiness gaps. Our benchmark is publicly available athttps://decodingtrust.github.io/; our dataset can be previewed athttps://huggingface.co/datasets/AI-Secure/DecodingTrust; a concise version ofthis work is at https://openreview.net/pdf?id=kaHpo8OZw2.

DecodingTrust: A Comprehensive Assessment of Trustworthiness in GPT Models