Unverified50% confidenceConceptExact time
Data contamination refers to the possibility that a model's training data already includes bug fix solutions publicly available on GitHub, meaning inference during testing may actually be memory recall.
1
Sources
50%
Confidence
Long-term
Relevance
7/19/2026
First Seen
Sources
Grok 4.5 vs Composer 2.5实测：真实代码库谁更强？
Related Claims
Unverified数据污染问题源于大模型在包含 GitHub、Stack Overflow、LeetCode 等平台公开代码题解的语料上进行预训练75% similarUnverified代码训练模型在构建训练集时会特别关注去重处理，因为GitHub上大量fork和复制代码会导致模型过度记忆特定代码片段而非学习通用模式71% similarUnverifiedGitHub Copilot通过海量真实代码补全的接受/拒绝反馈数据持续优化模型的代码生成质量，形成数据飞轮65% similarUnverifiedGitHub Copilot uses a Transformer architecture at its core and predicts code completions by learning the statistical patterns and semantic structures of code.65% similarUnverified暴露的.git目录可被攻击者利用git clone或专用工具（如GitTools）完整重建项目源码65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/561537API
curl https://kongchang.com/api/v1/knowledge/claims/561537MCP
get_claim(id=561537)