AutoRISE evolves red-teaming attack strategies as editable executable programs via an agent, yielding 17-point higher average attack success rates than baselines across 11 models.
--- DISCLOSURE (begin) --- Target response (excerpt): --- DISCLOSURE (begin) ---
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CR 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
AutoRISE: Agent-Driven Strategy Evolution for Red-Teaming Large Language Models
AutoRISE evolves red-teaming attack strategies as editable executable programs via an agent, yielding 17-point higher average attack success rates than baselines across 11 models.