Unverified50% confidenceConceptExact time
训练数据构成决定模型世界观边界的现象在数字人文领域被称为'语料库偏差'（Corpus Bias）
1
Sources
50%
Confidence
Long-term
Relevance
7/9/2026
First Seen
Sources
Related Claims
Unverifiedexposure bias的根源在于训练与推理的分布不匹配，训练时使用teacher forcing策略而推理时模型依赖自身生成的token73% similarUnverifiedIf training corpora contain gender, racial, or cultural bias, models will inherit these biases.71% similarUnverified来自不同厂商、经历不同训练路径的模型持有不同的先验偏见，从而在交叉审查中形成真正的视角差异71% similarUnverifiedThe gender bias in word embedding models originates from training corpora (such as the Google News corpus) reflecting pre-existing human social attitudes, rather than being artificially introduced.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/392316API
curl https://kongchang.com/api/v1/knowledge/claims/392316MCP
get_claim(id=392316)