Research map / Adversarial robustness and security / Attacks on text and graphs
LLM-Agnostic Semantic Representation Attack
Jiawei Lian, Jianhong Pan, Lefan Wang, Yi Wang, Tairan Huang, Shaohui Mei, Lap‐Pui Chau
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 |
|---|---|---|---|
| use Vicuna 7B as a semantic proxy model (SRA+ vs SRA) | SRA · HarmBench, Llama 2 7B | ASR (%): 30.33 -> 100.00 | Improved |
| use Vicuna 7B as a semantic proxy model (SRA+ vs SRA) | SRA · HarmBench, Llama 2 13B | ASR (%): 19.00 -> 100.00 | Improved |
| use Vicuna 7B as a semantic proxy model (SRA+ vs SRA) | SRA · HarmBench, Llama 2 70B | ASR (%): 17.33 -> 100.00 | Improved |
| use Vicuna 7B as a semantic proxy model (SRA+ vs SRA) | SRA · HarmBench, Qwen3 4B Thinking | ASR (%): 66.33 -> 100.00 | Improved |
| use Vicuna 7B as a semantic proxy model (SRA+ vs SRA) | SRA · HarmBench, Llama 3.1 8B | ASR (%): 45.00 -> 100.00 | Improved |
| use Vicuna 7B as a semantic proxy model (SRA+ vs SRA) | SRA · HarmBench, R2D2 7B | ASR (%): 42.00 -> 93.00 | Improved |
| use SRA instead of GCG | GCG · HarmBench, Llama 2 7B | ASR (%): 46.25 -> 30.33 | Hurt |
| replace GCG with SRA under 15s budget | GCG · AdvBench, Vicuna 7B | perplexity: 753.39 -> 24.21 | Improved |
| replace BEAST with SRA under 15s budget | BEAST · AdvBench, Vicuna 13B | ASR (%): 37.69 -> 86.73 | Improved |
| increase budget from 15s to 60s | SRA · AdvBench, Vicuna 7B | perplexity: 24.21 -> 18.67 | Improved |
| lower semantic threshold tau from 1000 to 10 (ablation) | SRA · Vicuna 13B, 15s budget | ASR (%): 75.38 -> 86.15 | Improved |
| lower semantic threshold tau from 100 to 20 (ablation) | SRA · Vicuna 13B, 15s budget | ASR (%): 77.69 -> 86.73 | Improved |
Closest papers by meaning
- Semantic Representation Attack against Aligned Large Language Models (NeurIPS 2025 · similarity 0.94)
- Exploiting LLM Quantization (NeurIPS 2024 · similarity 0.88)
- An LLM can Fool Itself: A Prompt-Based Adversarial Attack (ICLR 2024 · similarity 0.88)
- Soft Prompt Threats: Attacking Safety Alignment and Unlearning in Open-Source LLMs through the Embedding Space (NeurIPS 2024 · similarity 0.87)
- Safeguarding LLMs via Model-Agnostic Latent Safety Signals from Dark Knowledge (NeurIPS 2026 · similarity 0.87)
- A Frustratingly Simple Yet Highly Effective Attack Baseline: Over 90% Success Rate Against the Strong Black-box Models of GPT-4.5/4o/o1 (NeurIPS 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.
