Unverified50% confidenceFactExact time
闭源模型同样可能被越狱和被滥用，几乎所有主流闭源大模型都曾被用户通过越狱或提示词工程绕过安全护栏
1
Sources
50%
Confidence
Long-term
Relevance
8/7/2026
First Seen
Sources
Related Claims
UnverifiedExcessive blocking is itself a security risk because it prevents defenders from testing systems and releasing patches, while malicious actors can still use other models including open-source ones.74% similarVerifiedOnce open-source model weights are made public, attackers can remove a model's safety restrictions through fine-tuning.72% similarUnverifiedOnce model weights are publicly released, anyone can download, fine-tune, and remove safety alignment measures, rendering the traditional gatekeeper model ineffective.67% similarUnverifiedSTRIDE 模型（仿冒、篡改、抵赖、信息泄露、拒绝服务、权限提升）是威胁建模中最广泛使用的方法论之一66% similarUnverifiedThrough carefully crafted jailbreak prompts, attackers can bypass the safety guardrails of mainstream LLMs to generate harmful content.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/700093API
curl https://kongchang.com/api/v1/knowledge/claims/700093MCP
get_claim(id=700093)