Unverified50% confidenceFactExact time
Sensitive information from training data can be extracted by reverse-engineering model weights, which is known as a model inversion attack.
1
Sources
50%
Confidence
Long-term
Relevance
9/19/2026
First Seen
Sources
Related Entities
Related Claims
Unverified梯度反演攻击证明攻击者可通过分析上传的模型梯度高保真地重建原始训练图像甚至文本数据78% similarUnverifiedBackdoor attacks plant triggers in training data, causing the model to behave normally on regular inputs but produce preset malicious outputs when specific trigger words are encountered.75% similarVerifiedFine-tuning continues to update parameters on top of pre-trained weights through the backpropagation algorithm, adapting the model to a specific domain or task style.72% similarVerifiedFine-tuning involves taking an existing large model and further training it with a specially prepared proprietary dataset, thereby changing the model's weight values72% similarVerified当模型训练数据包含评测集题目或高度相似内容时，模型可能通过记忆而非真正理解获得高分72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/935771API
curl https://kongchang.com/api/v1/knowledge/claims/935771MCP
get_claim(id=935771)