Benchmark RadarAI BENCHMARK PROFILE
PPE-Bench
PPE-Bench evaluates machine unlearning in multimodal large language models under private-public entanglement, where images contain a target individual to forget and public elements to preserve.
- Released
- 2026-07-03
- Readiness
- Paper only
- Primary field
- General AI
Why it matters
Addresses the lack of benchmarks that reflect real-world image complexity and entanglement of private and public information, supporting evaluation of unlearning methods that must preserve public context.
Motivation
Multimodal Large Language Models (MLLMs) have shown strong capabilities, but they may memorize private information from web data, raising privacy concerns.
Primary resources
Benchmark Radar records only publicly supported details and links back to primary sources for verification.