INDICT: Code Generation with Internal Dialogues of Critiques for Both Security and Helpfulness

Hung Le, Doyen Sahoo, Yingbo Zhou, Caiming Xiong, Silvio Savarese

Keywords: code generation, safety, helpfulness, code security, large language model, critic, autonomous agent

TL;DR: We propose to improve code generation by both safety and helpfulness through a collaborative and autonomous system of dual tool-enhanced critics, providing knowledge-grounded critic feedbacks to support the LLM code generator.

Abstract: Large language models (LLMs) for code are typically trained to align with natural language instructions to closely follow their intentions and requirements. However, in many practical scenarios, it becomes increasingly challenging for these models to navigate the intricate boundary between helpfulness and safety, especially against highly complex yet potentially malicious instructions. In this work, we introduce INDICT: a new framework that empowers LLMs with Internal Dialogues of Critiques for both safety and helpfulness guidance. The internal dialogue is a dual cooperative system between a safety-driven critic and a helpfulness-driven critic. Each critic provides analysis against the given task and corresponding generated response, equipped with external knowledge queried through relevant code snippets and tools like web search and code interpreter. We engage the dual critic system in both code generation stage as well as code execution stage, providing preemptive and post-hoc guidance respectively to LLMs. We evaluated INDICT on 8 diverse tasks across 8 programming languages from 5 benchmarks, using LLMs from 7B to 70B parameters. We observed that our approach can provide an advanced level of critiques of both safety and helpfulness analysis, significantly improving the quality of output codes (+10% absolute improvements in all models).

Primary Area: Safety in machine learning

Submission Number: 15006