Research map / Large language models / Agents, code and math
DarkBench+: An Extended Benchmark for Evaluating Dark Patterns in Large Language Models
Yaowen Liu, Shenjia Jing, Yufei Wei, Shoumin Zhang, Jinglu Zhang, Zhen Mei, Liangliang Yue, Jiarui Wang, Peng Zhang
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 |
|---|---|---|---|
| enable thinking mode (ablation) | Qwen3-235B-A22B · DarkBench+ (Chinese) | avg trigger rate: 31.48 -> 29.12 | Improved |
| enable thinking mode (ablation) | Claude-Sonnet-4 · DarkBench+ (Chinese) | avg trigger rate: 20.07 -> 19.71 | No clear effect |
| enable thinking mode across models (ablation) | LLMs · DarkBench+ | avg trigger rate: 31.48 -> 29.12 (Qwen3-235B zh) | Mixed |
| scale Gemma3 from 4B to 12B (ablation) | Gamma3 · DarkBench+ (Chinese) | avg trigger rate: 36.88 -> 34.24 | Improved |
| scale Gemma3 from 12B to 27B (ablation) | Gamma3 · DarkBench+ (Chinese) | avg trigger rate: 34.24 -> 37.13 | Hurt |
| closed-source vs open-source models | LLMs · DarkBench+ | avg trigger rate: 31.4% -> 25.8% | Improved |
| evaluate in English instead of Chinese for Chinese-native models | Qwen3-235B-A22B · User Retention | trigger rate: 87.27 -> 96.33 | Hurt |
| model scale in Qwen3 series | Qwen3 · DarkBench+ | avg trigger rate | No clear effect |
Closest papers by meaning
- DarkBench: Benchmarking Dark Patterns in Large Language Models (ICLR 2025 · similarity 0.92)
- metabench - A Sparse Benchmark of Reasoning and Knowledge in Large Language Models (ICLR 2025 · similarity 0.88)
- CogBench: a large language model walks into a psychology lab (ICML 2024 · similarity 0.88)
- MDK12-Bench: A Multi-Discipline Benchmark for Evaluating Reasoning in Multimodal Large Language Models (AAAI 2026 · similarity 0.87)
- PHYBench: Holistic Evaluation of Physical Perception and Reasoning in Large Language Models (NeurIPS 2025 · similarity 0.87)
- RBench: Graduate-level Multi-disciplinary Benchmarks for LLM & MLLM Complex Reasoning Evaluation (ICML 2025 · similarity 0.87)
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.
