PrivaCI-Bench: Evaluating Privacy with Contextual Integrity and Legal Compliance (2025.acl-long)
Copied to clipboard
Haoran Li, Wenbin Hu, Huihao Jing, Yulin Chen, Qi Hu, Sirui Han, Tianshu Chu, Peizhao Hu, Yangqiu Song
| Challenge: | Recent advances in generative large language models (LLMs) have enabled wider applicability, accessibility, and flexibility. |
| Approach: | They propose a contextual privacy evaluation benchmark that covers the entire relevant social context through private information flows. |
| Outcome: | The proposed benchmarks cover legal compliance, real court cases, privacy policies, and synthetic data. |
Similar Papers
ContextLens: Modeling Imperfect Privacy and Safety Context for Legal Compliance (2026.acl-long)
Copied to clipboard
Haoran Li, Yulin Chen, Huihao Jing, Wenbin Hu, Tsz Ho Li, Chanhou Lou, Hong Ting Tsang, Sirui Han, Yangqiu Song
| Challenge: | Existing approaches to contextualize safety and privacy assessments assume the availability of complete and clear context, whereas real-world contexts tend to be ambiguous and incomplete. |
| Approach: | They propose a semi-rule-based framework that leverages large language models to ground the input context in the legal domain and explicitly identify both known and unknown factors for legal compliance. |
| Outcome: | The proposed framework can significantly improve existing baselines without training and can identify the ambiguous and missing factors. |
Context Reasoner: Incentivizing Reasoning Capability for Contextualized Privacy and Safety Compliance via Reinforcement Learning (2025.emnlp-main)
Copied to clipboard
Wenbin Hu, Haoran Li, Huihao Jing, Qi Hu, Ziqian Zeng, Sirui Han, Xu Heli, Tianshu Chu, Peizhao Hu, Yangqiu Song
| Challenge: | Current mitigation strategies fail to preserve contextual reasoning capabilities in risky scenarios, leading to systemic risks for legal compliance. |
| Approach: | They propose to use reinforcement learning with a rule-based reward to incentivize contextual reasoning capabilities while enhancing compliance with safety and privacy norms. |
| Outcome: | The proposed model outperforms Qwen2.5-7B-Instruct model in safety and privacy benchmarks and achieves +8.58% accuracy improvement. |
GoldCoin: Grounding Large Language Models in Privacy Laws via Contextual Integrity Theory (2024.emnlp-main)
Copied to clipboard
| Challenge: | Existing research studies privacy by exploring various privacy attacks, defenses, and evaluations within narrowly predefined patterns. |
| Approach: | They propose a framework that leverages the theory of contextual integrity as a bridge to help LLMs understand the complex contexts for judicial assessing privacy violations. |
| Outcome: | The proposed framework bridges the theory of contextual integrity as a bridge, creating numerous synthetic scenarios grounded in relevant privacy statutes (e.g., HIPAA). |
Privacy Checklist: Privacy Violation Detection Grounding on Contextual Integrity Theory (2025.naacl-long)
Copied to clipboard
| Challenge: | Existing privacy studies focus on sub-fields, but they focus on a few sub-domains. |
| Approach: | They propose to use the Health Insurance Portability and Accountability Act of 1996 as an example to develop a checklist that covers social identities, private attributes, and existing privacy regulations. |
| Outcome: | The proposed checklist covers social identities, private attributes, and existing privacy regulations. |
PrivLM-Bench: A Multi-level Privacy Evaluation Benchmark for Language Models (2024.acl-long)
Copied to clipboard
Haoran Li, Dadi Guo, Donghao Li, Wei Fan, Qi Hu, Xin Liu, Chunkit Chan, Duanyi Yao, Yuan Yao, Yangqiu Song
| Challenge: | generative large language models (LLMs) exhibit surprising capability and integrate previous tasks into a unified text generation formulation. |
| Approach: | They propose a privacy evaluation benchmark to quantify the privacy leakage of language models. |
| Outcome: | The proposed benchmark compares PPLMs with different privacy implementations to find out how privacy leakage is handled. |
PII-Bench: Evaluating Query-Aware Privacy Protection Systems (2026.acl-long)
Copied to clipboard
| Challenge: | Existing models do not detect PII in user prompts, despite their convenience . current models show significant limitations in determining PI I query relevance . |
| Approach: | They propose a query-unrelated PII masking strategy and propose PIi-Bench . they propose 'quick-and-easy' PI I masking with a user query and context description . |
| Outcome: | The proposed model performs well in basic PII detection, but shows significant limitations in query relevance. |
Does Context Matter? ContextualJudgeBench for Evaluating LLM-based Judges in Contextual Settings (2025.acl-long)
Copied to clipboard
| Challenge: | Contextual evaluation is challenging for state-of-the-art judge models . evaluation criteria are often conditional and dependent on practitioner priorities . |
| Approach: | They propose a judge benchmark that evaluates large language models as judges in contexts . they use human annotations and model-based perturbations to build the benchmark . |
| Outcome: | The proposed benchmark aims to evaluate large language models in contexts with 2,000 challenging response pairs. |
Pub-LawBench: Public-Oriented Benchmarking for LegalAI (2026.acl-long)
Copied to clipboard
| Challenge: | Existing evaluation frameworks focus on legal professionals, not legal professionals. |
| Approach: | They propose a public-oriented LegalAI benchmark grounded in legal functionalism and genre analysis to address this gap. |
| Outcome: | The proposed model evaluates 17 large language models on Pub-LawBench using simple prompts and Chain-of-Thought under a vanilla inference setting. |
PII-VisBench: Evaluating Personally Identifiable Information Safety in Vision Language Models Along a Continuum of Visibility (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing evaluations of PII leakage ignore how a subject’s online presence affects privacy alignment. |
| Approach: | They propose a benchmark that evaluates safety through the continuum of online presence by stratifying 200 subjects into four visibility categories: high, medium, low, and zero. |
| Outcome: | The proposed model stratifies 200 subjects into four visibility categories based on the extent and nature of their information available online. |
Keep Security! Benchmarking Security Policy Preservation in Large Language Model Contexts Against Indirect Attacks in Question Answering (2025.emnlp-main)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) are increasingly deployed in sensitive domains . large-scale benchmarks for contextual security preservation against attacks remain lacking . |
| Approach: | They evaluate 10 Large Language Models on a benchmark dataset to assess their adherence to contextual non-disclosure policies. |
| Outcome: | The proposed model fails to adhere to user-defined security policies in question answering . the model fails in indirect attacks, especially when it violates user-definable policies . |