SelfPrompt: Autonomously Evaluating LLM Robustness via Domain-Constrained Knowledge Guidelines and Refined Adversarial Prompts

TL;DR

This paper presents SelfPrompt, a framework for autonomous LLM robustness evaluation using domain-specific knowledge graphs and refined adversarial prompts, reducing reliance on traditional benchmarks and enabling targeted assessments.

Contribution

Introduces SelfPrompt, a novel self-evaluation framework that generates domain-constrained adversarial prompts for LLM robustness assessment without external benchmarks.

Findings

Effective in evaluating ChatGPT, Llama-3.1, Phi-3, and Mistral.

Reduces dependency on traditional evaluation datasets.

Provides targeted robustness insights in specific domains.

Abstract

Traditional methods for evaluating the robustness of large language models (LLMs) often rely on standardized benchmarks, which can escalate costs and limit evaluations across varied domains. This paper introduces a novel framework designed to autonomously evaluate the robustness of LLMs by incorporating refined adversarial prompts and domain-constrained knowledge guidelines in the form of knowledge graphs. Our method systematically generates descriptive sentences from domain-constrained knowledge graph triplets to formulate adversarial prompts, enhancing the relevance and challenge of the evaluation. These prompts, generated by the LLM itself and tailored to evaluate its own robustness, undergo a rigorous filtering and refinement process, ensuring that only those with high textual fluency and semantic fidelity are used. This self-evaluation mechanism allows the LLM to evaluate its robustness without the need for external benchmarks. We assess the effectiveness of our framework through extensive testing on both proprietary models like ChatGPT and open-source models such as Llama-3.1, Phi-3, and Mistral. Results confirm that our approach not only reduces dependency on conventional data but also provides a targeted and efficient means of evaluating LLM robustness in constrained domains.