Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces AdversaBench, an automated red-teaming pipeline that employs five structured operators to mutate seed prompts and validate adversarial failures through a multi-judge confirmation system. The study reveals that operator effectiveness varies significantly across categories, with instruction-following tasks requiring more iterations to produce confirmed failures compared to reasoning and tool-use tasks. Notably, adversarial prompts generated for one model successfully transfer to another model, indicating that the mutations exploit general behavioral patterns rather than specific vulnerabilities.
AdversaBench reveals that adversarial prompts can transfer across models, challenging the notion of model-specific weaknesses in LLMs.
Scaling adversarial evaluation of large language models requires both a method for generating hard inputs and a reliable way to confirm that resulting failures are real. We present AdversaBench, an end-to-end red-teaming pipeline that mutates seed prompts with five structured operators, queries a target model, and confirms failures through a three-judge panel with a meta-judge tiebreaker. We report experiments on 45 seeds across three categories: reasoning, instruction-following, and tool use. Every seed produced a confirmed failure. Four findings stand out. First, operator effectiveness varies sharply by category: inject_distractor scores 0.00 mean reward on instruction-following seeds but 0.80-0.83 on reasoning and tool-use. Second, binary failure rate hides difficulty: instruction-following seeds required 2.4 attacker iterations on average versus 1.1 for other categories, a gap visible in survival curves. Third, pairwise judge agreement of 80-87% coexists with near-zero Cohen's kappa due to label skew; category-level disagreement rates are more informative. Fourth, adversarial prompts generated against Llama 3.1 8B transfer zero-shot to Llama 3.3 70B, suggesting the mutations exploit general behavioral patterns rather than model-specific weaknesses. Code, dataset, and analysis scripts are available at https://github.com/khanak0509/AdversaBench .