# Red-Teaming Large Language Models using Chain of Utterances for Safety-Alignment

arXiv: https://arxiv.org/abs/2308.09662 · HTML: https://mlsanity.com/paper/2308.09662/

## Key facts

- Cited by 223 papers in the map (data through 14 Oct 2026).
- Placed in Prompt injection and agent security, one of 12 directions in LLM safety and security.
- Builds on 23 of its 34 references that are in the map.
- First posted on arXiv 18 Aug 2023.

## Placed in

- LLM safety and security / [Prompt injection and agent security](https://mlsanity.com/safety/prompt-injection-and-agent-security/)

## Builds on (most cited first, 23 of 23)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [LLaMA: Open and Efficient Foundation Language Models](https://mlsanity.com/paper/2302.13971/) (2023)
- [Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena](https://mlsanity.com/paper/2306.05685/) (2023)
- [Measuring Massive Multitask Language Understanding](https://mlsanity.com/paper/2009.03300/) (2020)
- [Large Language Models are Zero-Shot Reasoners](https://mlsanity.com/paper/2205.11916/) (2022)
- [Universal and Transferable Adversarial Attacks on Aligned Language Models](https://mlsanity.com/paper/2307.15043/) (2023)
- [Training Compute-Optimal Large Language Models](https://mlsanity.com/paper/2203.15556/) (2022)
- [TruthfulQA: Measuring How Models Mimic Human Falsehoods](https://mlsanity.com/paper/2109.07958/) (2021)
- [Challenging BIG-Bench Tasks and Whether Chain-of-Thought Can Solve Them](https://mlsanity.com/paper/2210.09261/) (2022)
- [A General Language Assistant as a Laboratory for Alignment](https://mlsanity.com/paper/2112.00861/) (2021)
- [Aligning AI With Shared Human Values](https://mlsanity.com/paper/2008.02275/) (2020)
- [BBQ: A Hand-Built Bias Benchmark for Question Answering](https://mlsanity.com/paper/2110.08193/) (2021)
- [Gender Bias in Coreference Resolution](https://mlsanity.com/paper/1804.09301/) (2018)
- [Principle-Driven Self-Alignment of Language Models from Scratch with Minimal Human Supervision](https://mlsanity.com/paper/2305.03047/) (2023)
- [On Second Thought, Let's Not Think Step by Step! Bias and Toxicity in Zero-Shot Reasoning](https://mlsanity.com/paper/2212.08061/) (2022)
- [Process for Adapting Language Models to Society (PALMS) with Values-Targeted Datasets](https://mlsanity.com/paper/2106.10328/) (2021)
- [The Capacity for Moral Self-Correction in Large Language Models](https://mlsanity.com/paper/2302.07459/) (2023)
- [Does Synthetic Data Generation of LLMs Help Clinical Text Mining?](https://mlsanity.com/paper/2303.04360/) (2023)
- [Legal Prompt Engineering for Multilingual Legal Judgement Prediction](https://mlsanity.com/paper/2212.02199/) (2022)
- [On the Educational Impact of ChatGPT: Is Artificial Intelligence Ready to Obtain a University Degree?](https://mlsanity.com/paper/2303.11146/) (2023)
- [Putting ChatGPT's Medical Advice to the (Turing) Test](https://mlsanity.com/paper/2301.10035/) (2023)
- [Beyond Classification: Financial Reasoning in State-of-the-Art Language Models](https://mlsanity.com/paper/2305.01505/) (2023)
- [Towards Safer Generative Language Models: A Survey on Safety Risks, Evaluations, and Improvements](https://mlsanity.com/paper/2302.09270/) (2023)

## Built on by (most cited first, 30 of 223)

- [SmoothLLM: Defending Large Language Models Against Jailbreaking Attacks](https://mlsanity.com/paper/2310.03684/) (2023)
- [LMSYS-Chat-1M: A Large-Scale Real-World LLM Conversation Dataset](https://mlsanity.com/paper/2309.11998/) (2023)
- [Shadow Alignment: The Ease of Subverting Safely-Aligned Language Models](https://mlsanity.com/paper/2310.02949/) (2023)
- [R-Judge: Benchmarking Safety Risk Awareness for LLM Agents](https://mlsanity.com/paper/2401.10019/) (2024)
- [Multilingual Jailbreak Challenges in Large Language Models](https://mlsanity.com/paper/2310.06474/) (2023)
- [Jailbreak Attacks and Defenses Against Large Language Models: A Survey](https://mlsanity.com/paper/2407.04295/) (2024)
- [SALAD-Bench: A Hierarchical and Comprehensive Safety Benchmark for Large Language Models](https://mlsanity.com/paper/2402.05044/) (2024)
- [Survey of Vulnerabilities in Large Language Models Revealed by Adversarial Attacks](https://mlsanity.com/paper/2310.10844/) (2023)
- [AEGIS: Online Adaptive AI Content Safety Moderation with Ensemble of LLM Experts](https://mlsanity.com/paper/2404.05993/) (2024)
- [Aegis2.0: A Diverse AI Safety Dataset and Risks Taxonomy for Alignment of LLM Guardrails](https://mlsanity.com/paper/2501.09004/) (2025)
- [Attacks, Defenses and Evaluations for LLM Conversation Safety: A Survey](https://mlsanity.com/paper/2402.09283/) (2024)
- [Large Language Models Often Know When They Are Being Evaluated](https://mlsanity.com/paper/2505.23836/) (2025)
- [Language Models are Homer Simpson! Safety Re-Alignment of Fine-tuned Language Models through Task Arithmetic](https://mlsanity.com/paper/2402.11746/) (2024)
- [Weak-to-Strong Jailbreaking on Large Language Models](https://mlsanity.com/paper/2401.17256/) (2024)
- [Aligner: Efficient Alignment by Learning to Correct](https://mlsanity.com/paper/2402.02416/) (2024)
- [Red-Teaming for Generative AI: Silver Bullet or Security Theater?](https://mlsanity.com/paper/2401.15897/) (2024)
- [FOFO: A Benchmark to Evaluate LLMs' Format-Following Capability](https://mlsanity.com/paper/2402.18667/) (2024)
- [SafetyPrompts: a Systematic Review of Open Datasets for Evaluating and Improving Large Language Model Safety](https://mlsanity.com/paper/2404.05399/) (2024)
- [JailbreakZoo: Survey, Landscapes, and Horizons in Jailbreaking Large Language and Vision-Language Models](https://mlsanity.com/paper/2407.01599/) (2024)
- [STAR-1: Safer Alignment of Reasoning LLMs with 1K Data](https://mlsanity.com/paper/2504.01903/) (2025)
- [The Art of Defending: A Systematic Evaluation and Analysis of LLM Defense Strategies on Safety and Over-Defensiveness](https://mlsanity.com/paper/2401.00287/) (2023)
- [Fight Back Against Jailbreaking via Prompt Adversarial Tuning](https://mlsanity.com/paper/2402.06255/) (2024)
- [Uncovering Safety Risks of Large Language Models through Concept Activation Vector](https://mlsanity.com/paper/2404.12038/) (2024)
- [T2VSafetyBench: Evaluating the Safety of Text-to-Video Generative Models](https://mlsanity.com/paper/2407.05965/) (2024)
- [Red Teaming Visual Language Models](https://mlsanity.com/paper/2401.12915/) (2024)
- [Navigating the Risks: A Survey of Security, Privacy, and Ethics Threats in LLM-Based Agents](https://mlsanity.com/paper/2411.09523/) (2024)
- [Aligning Large Language Models with Representation Editing: A Control Perspective](https://mlsanity.com/paper/2406.05954/) (2024)
- [Introducing v0.5 of the AI Safety Benchmark from MLCommons](https://mlsanity.com/paper/2404.12241/) (2024)
- [SG-Bench: Evaluating LLM Safety Generalization Across Diverse Tasks and Prompt Types](https://mlsanity.com/paper/2410.21965/) (2024)
- [Multi-Turn Context Jailbreak Attack on Large Language Models From First Principles](https://mlsanity.com/paper/2408.04686/) (2024)

## Abstract (arXiv)

Larger language models (LLMs) have taken the world by storm with their massive multi-tasking capabilities simply by optimizing over a next-word prediction objective. With the emergence of their properties and encoded knowledge, the risk of LLMs producing harmful outputs increases, making them unfit for scalable deployment for the public. In this work, we propose a new safety evaluation benchmark RED-EVAL that carries out red-teaming. We show that even widely deployed models are susceptible to the Chain of Utterances-based (CoU) prompting, jailbreaking closed source LLM-based systems such as GPT-4 and ChatGPT to unethically respond to more than 65% and 73% of harmful queries. We also demonstrate the consistency of the RED-EVAL across 8 open-source LLMs in generating harmful responses in more than 86% of the red-teaming attempts. Next, we propose RED-INSTRUCT--An approach for the safety alignment of LLMs. It constitutes two phases: 1) HARMFULQA data collection: Leveraging CoU prompting, we collect a dataset that consists of 1.9K harmful questions covering a wide range of topics, 9.5K safe and 7.3K harmful conversations from ChatGPT; 2) SAFE-ALIGN: We demonstrate how the conversational dataset can be used for the safety alignment of LLMs by minimizing the negative log-likelihood over helpful responses and penalizing over harmful responses by gradient accent over sample loss. Our model STARLING, a fine-tuned Vicuna-7B, is observed to be more safely aligned when evaluated on RED-EVAL and HHH benchmarks while preserving the utility of the baseline models (TruthfulQA, MMLU, and BBH).