Seamus_F

Message

Robustness of Contrast-Consistent Search to Adversarial Prompting

Produced as part of the AI Safety Hub Labs programme run by Charlie Griffin and Julia Karbing. This project was mentored by Nandi Schoots. Image generated by DALL-E 3. Introduction We look at how adversarial prompting affects the outputs of large language models (LLMs) and compare it with how the...

Nov 1, 2023•18

Message

11 karma

Member for 2 years

Seamus_F — LessWrong

Seamus_F

Message

Seamus_F

Robustness of Contrast-Consistent Search to Adversarial Prompting

Nov 1, 2023•18

Message

11 karma

Member for 2 years

Robustness of Contrast-Consistent Search to Adversarial Prompting

Nandi

Nandi, i, Jamie Wright, Seamus_F, hugofry+ 0 more

Nandi, i, Jamie Wright, Seamus_F, hugofry

Produced as part of the AI Safety Hub Labs programme run by Charlie Griffin and Julia Karbing. This project was mentored by Nandi Schoots.

Introduction

We look at how adversarial prompting affects the outputs of large language models (LLMs) and compare it with how the adversarial prompts affect Contrast-Consistent Search (CCS). Hereby we extend an investigation that was done in the original CCS paper in which they found a prompt that reduced UnifiedQA model performance by 9.5%.

We find several attacks that severely corrupt the question-answering abilities of LLMs, which we hope will be useful to other people. The approaches we take to manipulate the outputs of the models are to:... (read 1830 more words →)

LESSWRONG
LW

LESSWRONG
LW

Seamus_F

Seamus_F

Seamus_F

Robustness of Contrast-Consistent Search to Adversarial Prompting

Seamus_F

Seamus_F

Seamus_F

Robustness of Contrast-Consistent Search to Adversarial Prompting

Introduction