Code Red! On the Harmfulness of Applying Off-the-shelf Large Language Models to Programming Tasks
This addresses the critical issue of aligning LLMs with human values to prevent malicious misuse in programming tasks, providing a foundation for targeted alignment strategies in software engineering.
The paper tackles the problem of assessing the potential harmfulness of large language models (LLMs) in software engineering by developing a framework that includes a taxonomy, dataset, and automatic evaluator, finding significant disparities in alignment with some models like Openhermes being more harmful and larger models tending to be less harmful.
Nowadays, developers increasingly rely on solutions powered by Large Language Models (LLM) to assist them with their coding tasks. This makes it crucial to align these tools with human values to prevent malicious misuse. In this paper, we propose a comprehensive framework for assessing the potential harmfulness of LLMs within the software engineering domain. We begin by developing a taxonomy of potentially harmful software engineering scenarios and subsequently, create a dataset of prompts based on this taxonomy. To systematically assess the responses, we design and validate an automatic evaluator that classifies the outputs of a variety of LLMs both open-source and closed-source models, as well as general-purpose and code-specific LLMs. Furthermore, we investigate the impact of models size, architecture family, and alignment strategies on their tendency to generate harmful content. The results show significant disparities in the alignment of various LLMs for harmlessness. We find that some models and model families, such as Openhermes, are more harmful than others and that code-specific models do not perform better than their general-purpose counterparts. Notably, some fine-tuned models perform significantly worse than their base-models due to their design choices. On the other side, we find that larger models tend to be more helpful and are less likely to respond with harmful information. These results highlight the importance of targeted alignment strategies tailored to the unique challenges of software engineering tasks and provide a foundation for future work in this critical area.