待验证50% 置信事实精确时间
The early GPT-3 model demonstrated some code comprehension ability, but its training data was primarily natural language, limiting the accuracy and engineering practicality of its code generation.
1
来源数
50%
置信度
中期 (~90 天)
时效性
2026/7/2
首次发现
有效期至：2026/9/30
来源
Codex Hands-On Tutorial: A Complete Guide from Installation to Enterprise-Level AI Programming
bilibili图灵AI大模型2026/6/21
相关事实
待验证GPT-3 的预训练语料以自然语言为主，代码仅占约 1%，而 Codex 的微调数据集中代码占比超过 90%80% 相似已验证GPT-4、Claude 3、Gemini等现代大语言模型通过在海量代码语料库上预训练，已能理解自然语言描述并生成高质量可运行代码76% 相似待验证GPT-4、Claude等大语言模型在预训练阶段消化了海量软件工程文档、测试规范和历史缺陷报告，天然理解等价类划分、边界值分析等经典测试设计方法76% 相似待验证The concept of Natural Language Programming can be traced back to academic research, but only became practically viable after GPT-4-level models appeared.75% 相似
引用此条事实
Stable URI
https://kongchang.com/claim/41543API
curl https://kongchang.com/api/v1/knowledge/claims/41543MCP
get_claim(id=41543)