National University of Defense Technology proposes the JailExpert jailbreak framework
Start here
Researchers at the National University of Defense Technology published the jailbreak framework JailExpert on arXiv, which structures the results of historical jailbreak attacks into reusable experience: each experience records the initial malicious instruction, the complete jailbreak prompt, the jailbreak pattern, and the numbers of successes and failures. The framework uses text-embedding-3-small to compute the semantic drift between the instruction and the complete prompt, and after K-Means grouping selects from each group the experience with the highest occurrence frequency and historical success rate. Across 7 LLMs the average attack success rate reaches 90%, and attack efficiency (ASR-E) is 2.7 times that of the best baseline, ReNeLLM.
Generated by AI from reports · Updated 7 days ago
Attention and developments
Attention, hourly
Now 36, 0% in 24 hours·Peak 38 (Nov 6 03:00)
Related research
Papers whose work the reports mention
- The reports mention EasyJailbreakEasyJailbreak: A Unified Framework for Jailbreaking Large Language ModelsNational University of Defense Technology proposes the JailExpert jailbreak framework, with attack efficiency 2.7 times that of the best baseline
arXiv · 2024-03-18 · Deep reading
- The reports mention CodeChameleonCodeChameleon: Personalized Encryption Framework for Jailbreaking Large Language ModelsNational University of Defense Technology proposes the JailExpert jailbreak framework, with attack efficiency 2.7 times that of the best baseline
arXiv · 2024-02-26
Progress timeline
- WeChat · Weijiqi LixinNational University of Defense Technology proposes the JailExpert jailbreak framework, with attack efficiency 2.7 times that of the best baseline
Researchers at the National University of Defense Technology published the jailbreak framework JailExpert on arXiv, which structures historical jailbreak attack results into reusable experience, achieving an average attack success rate of 90% across 7 LLMs, with attack efficiency (ASR-E) 2.7 times that of the best baseline, ReNeLLM. Each experience record contains the initial malicious instruction, the complete jailbreak prompt, the jailbreak pattern, and the numbers of successes and failures; the framework uses text-embedding-3-small to compute the semantic drift between the instruction and the complete prompt, and after grouping via K-Means, selects from each group the pattern with the highest occurrence frequency and historical success rate as the representative template, then dynamically updates the counts after attacks. In the experiments, GPT-4's attack success rate was 76%, while Gemini-1.5-pro and GPT-4-Turbo reached 100%; LlamaGuard reduced GPT-4-Turbo's success rate from 95.5% to 62.4%, and PPL Filter had no effect on the attack success rate against Llama2-7b (97.3%).
Related discussion
Attention trend
New sources per day
- Nov 5: 1 new sources (1 outlets, 0 accounts)
- Nov 6: 0 new sources (0 outlets, 0 accounts)
- Nov 7: 0 new sources (0 outlets, 0 accounts)
- Nov 8: 0 new sources (0 outlets, 0 accounts)
- Nov 9: 0 new sources (0 outlets, 0 accounts)
- Nov 10: 0 new sources (0 outlets, 0 accounts)
- Nov 11: 0 new sources (0 outlets, 0 accounts)
- Nov 12: 0 new sources (0 outlets, 0 accounts)
- Nov 13: 0 new sources (0 outlets, 0 accounts)