Unverified50% confidenceFactExact time
2023年多项学术研究证明GPT-4、Claude等主流大模型均对提示注入攻击存在不同程度的脆弱性
1
Sources
50%
Confidence
Long-term
Relevance
7/6/2026
First Seen
Sources
Related Claims
UnverifiedResearch from institutions including ETH Zurich and Google DeepMind has demonstrated that top-tier models such as GPT-4 and Claude exhibit significant vulnerability to adversarial prompts, with success rates sometimes exceeding 50%.81% similarUnverifiedResearch shows that top models such as GPT-4 and Claude sometimes have a success rate exceeding 50% under carefully crafted adversarial attacks.81% similarUnverified该论文通过在提示词末尾附加自动生成的对抗性字符串，成功让GPT-4、Claude等多个顶级模型产生有害输出，且该攻击可跨模型迁移79% similarUnverifiedDifferent large models (GPT-4, Claude, Llama, etc.) differ in their response characteristics to prompts, and in practice it is necessary to fine-tune for specific models.75% similarUnverifiedResearchers have identified statistical anomalies indicative of benchmark contamination in mainstream models such as GPT-4 and Claude.75% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/113189API
curl https://kongchang.com/api/v1/knowledge/claims/113189MCP
get_claim(id=113189)