# Make Them Spill the Beans! Coercive Knowledge Extraction from (Production) LLMs

arXiv: https://arxiv.org/abs/2312.04782 · HTML: https://mlsanity.com/paper/2312.04782/

## Key facts

- Cited by 27 papers in the map (data through 14 Oct 2026).
- Builds on 46 of its 55 references that are in the map.
- First posted on arXiv 8 Dec 2023.

## Builds on (most cited first, 30 of 46)

- [GPT-4 Technical Report](https://mlsanity.com/paper/2303.08774/) (2023)
- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [LLaMA: Open and Efficient Foundation Language Models](https://mlsanity.com/paper/2302.13971/) (2023)
- [RoBERTa: A Robustly Optimized BERT Pretraining Approach](https://mlsanity.com/paper/1907.11692/) (2019)
- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [Chain-of-Thought Prompting Elicits Reasoning in Large Language Models](https://mlsanity.com/paper/2201.11903/) (2022)
- [Large Language Models are Zero-Shot Reasoners](https://mlsanity.com/paper/2205.11916/) (2022)
- [Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback](https://mlsanity.com/paper/2204.05862/) (2022)
- [Universal and Transferable Adversarial Attacks on Aligned Language Models](https://mlsanity.com/paper/2307.15043/) (2023)
- [A Survey of Large Language Models](https://mlsanity.com/paper/2303.18223/) (2023)
- [G-Eval: NLG Evaluation using GPT-4 with Better Human Alignment](https://mlsanity.com/paper/2303.16634/) (2023)
- [Code Llama: Open Foundation Models for Code](https://mlsanity.com/paper/2308.12950/) (2023)
- [Jailbroken: How Does LLM Safety Training Fail?](https://mlsanity.com/paper/2307.02483/) (2023)
- [Jailbreaking Black Box Large Language Models in Twenty Queries](https://mlsanity.com/paper/2310.08419/) (2023)
- [Fine-tuning Aligned Language Models Compromises Safety, Even When Users Do Not Intend To!](https://mlsanity.com/paper/2310.03693/) (2023)
- [Machine Unlearning](https://mlsanity.com/paper/1912.03817/) (2019)
- [BeaverTails: Towards Improved Safety Alignment of LLM via a Human-Preference Dataset](https://mlsanity.com/paper/2307.04657/) (2023)
- [ZOO: Zeroth Order Optimization based Black-box Attacks to Deep Neural Networks without Training Substitute Models](https://mlsanity.com/paper/1708.03999/) (2017)
- [GPTFUZZER: Red Teaming Large Language Models with Auto-Generated Jailbreak Prompts](https://mlsanity.com/paper/2309.10253/) (2023)
- [Improving alignment of dialogue agents via targeted human judgements](https://mlsanity.com/paper/2209.14375/) (2022)
- [Jailbreaking ChatGPT via Prompt Engineering: An Empirical Study](https://mlsanity.com/paper/2305.13860/) (2023)
- ["Do Anything Now": Characterizing and Evaluating In-The-Wild Jailbreak Prompts on Large Language Models](https://mlsanity.com/paper/2308.03825/) (2023)
- [Spectral Signatures in Backdoor Attacks](https://mlsanity.com/paper/1811.00636/) (2018)
- [Detecting Backdoor Attacks on Deep Neural Networks by Activation Clustering](https://mlsanity.com/paper/1811.03728/) (2018)
- [GPT-4 Is Too Smart To Be Safe: Stealthy Chat with LLMs via Cipher](https://mlsanity.com/paper/2308.06463/) (2023)
- [Is ChatGPT a Good NLG Evaluator? A Preliminary Study](https://mlsanity.com/paper/2303.04048/) (2023)
- [Remember What You Want to Forget: Algorithms for Machine Unlearning](https://mlsanity.com/paper/2103.03279/) (2021)
- [DeepInception: Hypnotize Large Language Model to Be Jailbreaker](https://mlsanity.com/paper/2311.03191/) (2023)
- [A Deep Reinforced Model for Abstractive Summarization](https://mlsanity.com/paper/1705.04304/) (2017)
- [Hard Prompts Made Easy: Gradient-Based Discrete Optimization for Prompt Tuning and Discovery](https://mlsanity.com/paper/2302.03668/) (2023)

## Built on by (most cited first, 27 of 27)

- [Great, Now Write an Article About That: The Crescendo Multi-Turn LLM Jailbreak Attack](https://mlsanity.com/paper/2404.01833/) (2024)
- [Jailbreak Attacks and Defenses Against Large Language Models: A Survey](https://mlsanity.com/paper/2407.04295/) (2024)
- [Keeping LLMs Aligned After Fine-tuning: The Crucial Role of Prompt Templates](https://mlsanity.com/paper/2402.18540/) (2024)
- [RigorLLM: Resilient Guardrails for Large Language Models against Undesired Content](https://mlsanity.com/paper/2403.13031/) (2024)
- [SoK: Membership Inference Attacks on LLMs are Rushing Nowhere (and How to Fix It)](https://mlsanity.com/paper/2406.17975/) (2024)
- [Operationalizing a Threat Model for Red-Teaming Large Language Models (LLMs)](https://mlsanity.com/paper/2407.14937/) (2024)
- [Immune: Improving Safety Against Jailbreaks in Multi-modal LLMs via Inference-Time Alignment](https://mlsanity.com/paper/2411.18688/) (2024)
- [Alpaca against Vicuna: Using LLMs to Uncover Memorization of LLMs](https://mlsanity.com/paper/2403.04801/) (2024)
- [JailbreakEval: An Integrated Toolkit for Evaluating Jailbreak Attempts Against Large Language Models](https://mlsanity.com/paper/2406.09321/) (2024)
- [AdvPrefix: An Objective for Nuanced LLM Jailbreaks](https://mlsanity.com/paper/2412.10321/) (2024)
- [The VLLM Safety Paradox: Dual Ease in Jailbreak Attack and Defense](https://mlsanity.com/paper/2411.08410/) (2024)
- [Rapid Optimization for Jailbreaking LLMs via Subconscious Exploitation and Echopraxia](https://mlsanity.com/paper/2402.05467/) (2024)
- [An Embarrassingly Simple Defense Against LLM Abliteration Attacks](https://mlsanity.com/paper/2505.19056/) (2025)
- [NeuRel-Attack: Neuron Relearning for Safety Disalignment in Large Language Models](https://mlsanity.com/paper/2504.21053/) (2025)
- [ASTRA: Autonomous Spatial-Temporal Red-teaming for AI Software Assistants](https://mlsanity.com/paper/2508.03936/) (2025)
- [LIAR: Leveraging Inference Time Alignment (Best-of-N) to Jailbreak LLMs in Seconds](https://mlsanity.com/paper/2412.05232/) (2024)
- [JULI: Jailbreak Large Language Models by Self-Introspection](https://mlsanity.com/paper/2505.11790/) (2025)
- [LLMs Can Unlearn Refusal with Only 1,000 Benign Samples](https://mlsanity.com/paper/2601.19231/) (2026)
- [Safety Recovery in Reasoning Models Is Only a Few Early Steering Steps Away](https://mlsanity.com/paper/2602.11096/) (2026)
- [SlotGCG: Exploiting the Positional Vulnerability in LLMs for Jailbreak Attacks](https://mlsanity.com/paper/2606.05609/) (2026)
- [SoK: A Comprehensive Security Analysis of Jailbreak Resilience in GPT and DeepSeek Models](https://mlsanity.com/paper/2506.18543/) (2025)
- [Layer-Wise Perturbations via Sparse Autoencoders for Adversarial Text Generation](https://mlsanity.com/paper/2508.10404/) (2025)
- [NeuroBreak: Unveil Internal Jailbreak Mechanisms in Large Language Models](https://mlsanity.com/paper/2509.03985/) (2025)
- [SABER: Uncovering Vulnerabilities in Safety Alignment via Cross-Layer Residual Connection](https://mlsanity.com/paper/2509.16060/) (2025)
- [Innocence in the Crossfire: Roles of Skip Connections in Jailbreaking Visual Language Models](https://mlsanity.com/paper/2507.13761/) (2025)
- [Matching Ranks Over Probability Yields Truly Deep Safety Alignment](https://mlsanity.com/paper/2512.05518/) (2025)
- [A Causal Perspective for Enhancing Jailbreak Attack and Defense](https://mlsanity.com/paper/2602.04893/) (2026)

## Abstract (arXiv)

Large Language Models (LLMs) are now widely used in various applications, making it crucial to align their ethical standards with human values. However, recent jail-breaking methods demonstrate that this alignment can be undermined using carefully constructed prompts. In our study, we reveal a new threat to LLM alignment when a bad actor has access to the model's output logits, a common feature in both open-source LLMs and many commercial LLM APIs (e.g., certain GPT models). It does not rely on crafting specific prompts. Instead, it exploits the fact that even when an LLM rejects a toxic request, a harmful response often hides deep in the output logits. By forcefully selecting lower-ranked output tokens during the auto-regressive generation process at a few critical output positions, we can compel the model to reveal these hidden responses. We term this process model interrogation. This approach differs from and outperforms jail-breaking methods, achieving 92% effectiveness compared to 62%, and is 10 to 20 times faster. The harmful content uncovered through our method is more relevant, complete, and clear. Additionally, it can complement jail-breaking strategies, with which results in further boosting attack performance. Our findings indicate that interrogation can extract toxic knowledge even from models specifically designed for coding tasks.