Unverified50% confidenceFactExact time
对抗训练由Goodfellow等人在对抗样本研究中系统化提出，通过在训练中引入对抗样本使模型学会识别和抵抗攻击
1
Sources
50%
Confidence
Long-term
Relevance
7/17/2026
First Seen
Sources
Related Claims
UnverifiedAdversarial training improves a model's ability to recognize and resist attack prompts by incorporating known jailbreak samples into training data, but faces the ongoing challenge of novel variants.83% similarVerifiedTraining data poisoning can implant persistent backdoors at the model weight level, posing far greater risks than application-layer jailbreaks, but the implementation threshold is higher and typically requires control over the training pipeline.68% similarUnverified若模型在训练过程中意识到自身会被审计，理论上存在通过学习规避审计特征来通过检测的可能，这一风险在具有强化学习背景的模型中尤为值得关注66% similarUnverifiedApplication-layer jailbreak defenses must simultaneously rely on robustness improvements at the training stage and dynamic monitoring at the inference stage, as reinforcement along a single dimension yields limited results.65% similarUnverifiedAlignment training optimizes the model's statistical behavior in known scenarios rather than establishing formally verifiable security boundaries, and may fail when faced with rare instruction combinations or role-playing frameworks outside the training distribution.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/544328API
curl https://kongchang.com/api/v1/knowledge/claims/544328MCP
get_claim(id=544328)