NeurIPS 2025 Explorer
Concepts
Authors
Glossary
johnsanterre.github.io
security evaluation
3 papers
AgentAuditor: Human-level Safety and Security Evaluation for LLM Agents
SEC-bench: Automated Benchmarking of LLM Agents on Real-World Software Security Tasks
Transferable Black-Box One-Shot Forging of Watermarks via Image Preference Models