REVIEW 7 cited by
Does Refusal Training in LLMs Generalize to the Past Tense?
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
read the original abstract
Refusal training is widely used to prevent LLMs from generating harmful, undesirable, or illegal outputs. We reveal a curious generalization gap in the current refusal training approaches: simply reformulating a harmful request in the past tense (e.g., "How to make a Molotov cocktail?" to "How did people make a Molotov cocktail?") is often sufficient to jailbreak many state-of-the-art LLMs. We systematically evaluate this method on Llama-3 8B, Claude-3.5 Sonnet, GPT-3.5 Turbo, Gemma-2 9B, Phi-3-Mini, GPT-4o mini, GPT-4o, o1-mini, o1-preview, and R2D2 models using GPT-3.5 Turbo as a reformulation model. For example, the success rate of this simple attack on GPT-4o increases from 1% using direct requests to 88% using 20 past tense reformulation attempts on harmful requests from JailbreakBench with GPT-4 as a jailbreak judge. Interestingly, we also find that reformulations in the future tense are less effective, suggesting that refusal guardrails tend to consider past historical questions more benign than hypothetical future questions. Moreover, our experiments on fine-tuning GPT-3.5 Turbo show that defending against past reformulations is feasible when past tense examples are explicitly included in the fine-tuning data. Overall, our findings highlight that the widely used alignment techniques -- such as SFT, RLHF, and adversarial training -- employed to align the studied models can be brittle and do not always generalize as intended. We provide code and jailbreak artifacts at https://github.com/tml-epfl/llm-past-tense.
Forward citations
Cited by 7 Pith papers
-
SoK: Systematizing LLM Prompt Security: Taxonomies, Datasets, and Unified Evaluation of Attacks and Defenses
A systemization of LLM jailbreak security that adds linked taxonomies, an evaluation platform, and JailbreakDB, while its main attack–defense comparison results remain deferred.
-
HauntAttack: When Attack Follows Reasoning as a Shadow
HauntAttack embeds harmful instructions into reasoning-question conditions and reports a 70% average attack success rate across 11 large reasoning models, outperforming prior jailbreak baselines.
-
Cyber-Capable AI Agents: Vulnerabilities, Evaluation Containment, and Defensive Response
A structured review organizes cyber-capable-agent risks into five vulnerability classes and argues that evaluation environments must be treated as operational security systems rather than background.
-
Layer-Wise Perturbations via Sparse Autoencoders for Adversarial Text Generation
Sparse autoencoder activation perturbation (SFPF) applied on top of existing jailbreak prompts raises attack success rate on Qwen3-32B, but with no defense evaluation and weak reproducibility.
-
AlphaAlign: Incentivizing Safety Alignment with Extremely Simplified Reinforcement Learning
AlphaAlign uses pure reinforcement learning with a verifiable safety reward to make LLMs refuse harmful requests with explicit reasoning while preserving helpfulness on benign queries.
-
InfoFlood: Jailbreaking Large Language Models with Information Overload
InfoFlood claims near-perfect jailbreak success on four frontier LLMs by rewriting harmful queries into verbose academic prose with fake citations, past-tense framing, and ethical disclaimers, without adversarial suffixes.
-
A Multimodal Automatic Redteaming Evaluation based on Atomic Jailbreak Strategy Decoupling and Combination
A new jailbreak framework, HACA, combines atomic text and image attack strategies selected by a cross-modal planner and generates attacks with LLMs and text-to-image models, reaching 95.48% average attack success acro...
Discussion (0). Sign in to comment.