QuasiLeak-Bench: Evaluating LLM-Based Privacy Detectors on Policy-Conditioned Quasi-Evidence Group Recovery
We introduce QuasiLeak-Bench, a benchmark for policy-conditioned quasi-evidence group recovery, which evaluates whether detectors recover complete quasi-evidence groups and flag the disclosures the given policy prohibits.
ProofPaper ↗
Key points
- Natural language messages often contain quasi-identifiers that are weak in isolation but can jointly reveal an individual's private attributes.
- We call such a jointly revealing set of cues a quasi-evidence group.
- Existing privacy benchmarks focus on explicit identifiers, attribute inference, or task-driven disclosures, leaving inference from quasi-evidence groups largely unmeasured.
- In practice, whether exposing such a quasi-evidence group violates privacy also depends on sender-specified disclosure policies.
Sources (1)
- [1]QuasiLeak-Bench: Evaluating LLM-Based Privacy Detectors on Policy-Conditioned Quasi-Evidence Group RecoveryarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 15, 06:20 AM
We introduce QuasiLeak-Bench, a benchmark for policy-conditioned quasi-evidence group recovery, which evaluates whether detectors recover complete quasi-evidence groups and flag the disclosures the given policy prohibits.
Natural language messages often contain quasi-identifiers that are weak in isolation but can jointly reveal an individual's private attributes.
Extractive summary: sentences quoted from the sources.
Before this
- Oct 15, 2026VALCON: Benchmarking and Enhancing Value Consistency in Large Language Models via Dual-Role Variation
- Oct 15, 2026MemScope: Evolving Agent Memory Architectures through Localized Code Repair
- Oct 15, 2026INCONVENIENT: A Benchmark to Assess LLM Generalization Capabilities from Convenience Samples
- Oct 14, 2026mistralai/Timestral-209M
- Oct 14, 2026From Failures to Learnable Worlds: Policy-Conditioned Environment Generation for Navigation Agents
- Oct 13, 2026Gemini 4 Argon for Robotics