Unverified50% confidenceOpinionExact time
Pretraining backdoors differ from common fine-tuning poisoning: the attack occurs in the early stage when the model learns general language capabilities, and once poisoned data is mixed into a massive corpus it is extremely difficult to audit item by item, and subsequent safety alignment and fine-tuning often cannot thoroughly remove it.
1
Sources
50%
Confidence
Long-term
Relevance
10/31/2026
First Seen
Sources
OpenCode Responds to Model Weakening Controversy; Microsoft Open-Sources 4B Desktop Agent
bilibili阿梨Aria早鸟报10/31/2026
Related Entities
Related Claims
UnverifiedData poisoning occurs during the training or fine-tuning stage, where attackers mix malicious samples into the dataset, causing the model to produce erroneous behavior under specific trigger conditions.76% similarUnverified2023年多项学术研究证明，即使经过严格安全对齐训练的模型，当恶意指令以自然语言文档形式出现在上下文窗口中时，其遵循率仍显著高于直接注入场景74% similarUnverifiedCybersecurity capability filtering can often only reduce the model's performance under specific prompts and can hardly erase knowledge already learned in the pretraining stage.74% similarVerifiedBenchmark data contamination refers to the phenomenon where, after a test set is made public, its questions may appear in the pre-training corpora of subsequent models, causing scores to reflect memorization rather than genuine reasoning ability.74% similarUnverifiedThere is an inherent tension in dangerous capability assessments between the coverage, update frequency, and confidentiality of test sets: overly public test sets become incorporated into training data and lose effectiveness, while overly confidential tests lack peer-review credibility.73% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1120162API
curl https://kongchang.com/api/v1/knowledge/claims/1120162MCP
get_claim(id=1120162)