ASTPrompter: Weakly Supervised Automated Language Model Red-Teaming to Identify Low-Perplexity Toxic Prompts

• URL: http://arxiv.org/abs/2407.09447v3
• Date: Fri, 31 Jan 2025 08:22:46 GMT
• Title: ASTPrompter: Weakly Supervised Automated Language Model Red-Teaming to Identify Low-Perplexity Toxic Prompts
• Authors: Amelia F. Hardy, Houjun Liu, Bernard Lange, Duncan Eddy, Mykel J. Kochenderfer,
• Abstract summary: We propose a reinforcement learning formulation of red-teaming to discover prompts that both elicit toxic outputs from a defender and have low perplexity as scored by that defender. Our policy performs competitively, producing prompts that induce defender toxicity at a rate of 2-23 times higher than baseline across model scales. Our method creates black-box attacks with 5.4-14 times increased toxicity.
• Score: 31.481630330369427
• License:
• Abstract: Conventional approaches for the automated red-teaming of large language models (LLMs) aim to identify prompts that elicit toxic outputs from a frozen language model (the defender). This often results in the prompting model (the adversary) producing text that is unlikely to arise during autoregression. In response, we propose a reinforcement learning formulation of LLM red-teaming designed to discover prompts that both (1) elicit toxic outputs from a defender and (2) have low perplexity as scored by that defender. These prompts are the most pertinent in a red-teaming setting because the defender generates them with high probability. We solve this formulation with an online and weakly supervised form of Identity Preference Optimization (IPO), attacking models ranging from 137M to 7.8B parameters. Our policy performs competitively, producing prompts that induce defender toxicity at a rate of 2-23 times higher than baseline across model scales. Importantly, these prompts have lower perplexity than both automatically generated and human-written attacks. Furthermore, our method creates black-box attacks with 5.4-14 times increased toxicity. To assess the downstream utility of our method, we use rollouts from our policy as negative examples for downstream toxicity tuning and demonstrate improved safety.

Related papers

• Turning Logic Against Itself : Probing Model Defenses Through Contrastive Questions [51.51850981481236]
  We introduce POATE, a novel jailbreak technique that harnesses contrastive reasoning to provoke unethical responses. PoATE crafts semantically opposing intents and integrates them with adversarial templates, steering models toward harmful outputs with remarkable subtlety. To counter this, we propose Intent-Aware CoT and Reverse Thinking CoT, which decompose queries to detect malicious intent and reason in reverse to evaluate and reject harmful responses.
  arXiv  Detail & Related papers  (2025-01-03T15:40:03Z)
• The Dark Side of Human Feedback: Poisoning Large Language Models via User Inputs [8.449922248196705]
  We present a subtle yet effective poisoning attack via user-supplied prompts to penetrate alignment training protections. Our attack, even without explicit knowledge about the target LLMs in the black-box setting, subtly alters the reward feedback mechanism. By injecting 1% of these specially crafted prompts into the data, through malicious users, we demonstrate a toxicity score up to two times higher when a specific trigger word is used.
  arXiv  Detail & Related papers  (2024-09-01T17:40:04Z)
• Learning diverse attacks on large language models for robust red-teaming and safety tuning [126.32539952157083]
  Red-teaming, or identifying prompts that elicit harmful responses, is a critical step in ensuring the safe deployment of large language models. We show that even with explicit regularization to favor novelty and diversity, existing approaches suffer from mode collapse or fail to generate effective attacks. We propose to use GFlowNet fine-tuning, followed by a secondary smoothing phase, to train the attacker model to generate diverse and effective attack prompts.
  arXiv  Detail & Related papers  (2024-05-28T19:16:17Z)
• Tiny Refinements Elicit Resilience: Toward Efficient Prefix-Model Against LLM Red-Teaming [37.32997502058661]
  This paper introduces the textbfsentinel model as a plug-and-play prefix module designed to reconstruct the input prompt with just a few tokens. The sentinel model naturally overcomes the textit parameter inefficiency and textitlimited model accessibility for fine-tuning large target models. Our experiments across text-to-text and text-to-image demonstrate the effectiveness of our approach in mitigating toxic outputs.
  arXiv  Detail & Related papers  (2024-05-21T08:57:44Z)
• Towards Building a Robust Toxicity Predictor [13.162016701556725]
  This paper presents a novel adversarial attack, texttToxicTrap, introducing small word-level perturbations to fool SOTA text classifiers to predict toxic text samples as benign. Two novel goal function designs allow ToxicTrap to identify weaknesses in both multiclass and multilabel toxic language detectors.
  arXiv  Detail & Related papers  (2024-04-09T22:56:05Z)
• DPP-Based Adversarial Prompt Searching for Lanugage Models [56.73828162194457]
  Auto-regressive Selective Replacement Ascent (ASRA) is a discrete optimization algorithm that selects prompts based on both quality and similarity with determinantal point process (DPP) Experimental results on six different pre-trained language models demonstrate the efficacy of ASRA for eliciting toxic content.
  arXiv  Detail & Related papers  (2024-03-01T05:28:06Z)
• Unveiling the Implicit Toxicity in Large Language Models [77.90933074675543]
  The open-endedness of large language models (LLMs) combined with their impressive capabilities may lead to new safety issues when being exploited for malicious use. We show that LLMs can generate diverse implicit toxic outputs that are exceptionally difficult to detect via simply zero-shot prompting. We propose a reinforcement learning (RL) based attacking method to further induce the implicit toxicity in LLMs.
  arXiv  Detail & Related papers  (2023-11-29T06:42:36Z)
• AutoDAN: Interpretable Gradient-Based Adversarial Attacks on Large Language Models [55.748851471119906]
  Safety alignment of Large Language Models (LLMs) can be compromised with manual jailbreak attacks and (automatic) adversarial attacks. Recent studies suggest that defending against these attacks is possible: adversarial attacks generate unlimited but unreadable gibberish prompts, detectable by perplexity-based filters. We introduce AutoDAN, an interpretable, gradient-based adversarial attack that merges the strengths of both attack types.
  arXiv  Detail & Related papers  (2023-10-23T17:46:07Z)
• Fooling the Textual Fooler via Randomizing Latent Representations [13.77424820701913]
  adversarial word-level perturbations are well-studied and effective attack strategies. We propose a lightweight and attack-agnostic defense whose main goal is to perplex the process of generating an adversarial example. We empirically demonstrate near state-of-the-art robustness of AdvFooler against representative adversarial word-level attacks.
  arXiv  Detail & Related papers  (2023-10-02T06:57:25Z)
• TASA: Deceiving Question Answering Models by Twin Answer Sentences Attack [93.50174324435321]
  We present Twin Answer Sentences Attack (TASA), an adversarial attack method for question answering (QA) models. TASA produces fluent and grammatical adversarial contexts while maintaining gold answers.
  arXiv  Detail & Related papers  (2022-10-27T07:16:30Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.