Flames: Benchmarking Value Alignment of LLMs in Chinese

TL;DR

The paper introduces Flames, a comprehensive benchmark for evaluating Chinese LLMs' alignment with human values, revealing current models' safety gaps and proposing a new evaluation tool.

Contribution

It presents Flames, a novel benchmark incorporating Chinese cultural values and complex adversarial prompts, to better assess LLMs' safety and moral alignment.

Findings

All evaluated LLMs perform poorly on Flames, especially in safety and fairness.

Existing benchmarks are insufficient for uncovering safety vulnerabilities.

A lightweight scoring method enables efficient multi-dimensional evaluation.

Abstract

The widespread adoption of large language models (LLMs) across various regions underscores the urgent need to evaluate their alignment with human values. Current benchmarks, however, fall short of effectively uncovering safety vulnerabilities in LLMs. Despite numerous models achieving high scores and 'topping the chart' in these evaluations, there is still a significant gap in LLMs' deeper alignment with human values and achieving genuine harmlessness. To this end, this paper proposes a value alignment benchmark named Flames, which encompasses both common harmlessness principles and a unique morality dimension that integrates specific Chinese values such as harmony. Accordingly, we carefully design adversarial prompts that incorporate complex scenarios and jailbreaking methods, mostly with implicit malice. By prompting 17 mainstream LLMs, we obtain model responses and rigorously annotate them for detailed evaluation. Our findings indicate that all the evaluated LLMs demonstrate relatively poor performance on Flames, particularly in the safety and fairness dimensions. We also develop a lightweight specified scorer capable of scoring LLMs across multiple dimensions to efficiently evaluate new models on the benchmark. The complexity of Flames has far exceeded existing benchmarks, setting a new challenge for contemporary LLMs and highlighting the need for further alignment of LLMs. Our benchmark is publicly available at https://github.com/AIFlames/Flames.