Research map / Large language models / AI and society
HEV Generative Sandbox: A Framework for Assessing Domain-Specific Social Risks Through Human-LLM Simulation
Yiran Liu, Zhiyi Hou, Xiaoang Xu, Shuo Wang, Huijia Wu, Kaicheng Yu, Yang Yu, ChengXiang Zhai
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| switch from Secure-Baseline to Adversarial user | 7 LLMs average · HEV Sandbox all domains | Loss Event Frequency (%): 12.7 -> 55.0 | Hurt |
| switch from Standard to Perturbation user | 7 LLMs average · HEV Sandbox all domains | Loss Event Frequency (%): 16.6 -> 17.8 | No clear effect |
| switch from Standard to Perturbation user | Qwen-3-8B · HEV Sandbox all domains | Loss Event Frequency (%): 22.7 -> 23.5 | Hurt |
| move from encyclopedic to legal domain | 7 LLMs average · HEV Sandbox (Standard User) | Loss Event Frequency (%): 9.0 -> 28.2 | Hurt |
| use secure-baseline prompts instead of standard prompts | Vicuna-7B · HEV Sandbox all domains | Loss Event Frequency (%): 17.9 -> 15.7 | Improved |
| use closed-source instead of open-source models under adversarial user | GPT-4/DeepSeek/Zhipu vs GLM-4/Llama-3 · HEV Sandbox all domains | Loss Event Frequency (%): 41.5-43.7 -> 60.1-62.0 | Hurt |
Closest papers by meaning
- SafetyPrompts: A Systematic Review of Open Datasets for Evaluating and Improving Large Language Model Safety (AAAI 2025 · similarity 0.87)
- ALI-Agent: Assessing LLMs' Alignment with Human Values via Agent-based Evaluation (NeurIPS 2024 · similarity 0.87)
- Position: LLM Social Simulations Are a Promising Research Method (ICML 2025 · similarity 0.87)
- CASE-Bench: Context-Aware SafEty Benchmark for Large Language Models (ICML 2025 · similarity 0.86)
- SORRY-Bench: Systematically Evaluating Large Language Model Safety Refusal (ICLR 2025 · similarity 0.86)
- Navigating the Safety Landscape: Measuring Risks in Finetuning Large Language Models (NeurIPS 2024 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
