Verified80% confidenceFactTime unknown
GLM采用了独特的自回归填空预训练范式，与GPT的纯自回归和BERT的纯掩码语言模型都有所不同
10
Sources
80%
Confidence
Long-term
Relevance
6/1/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedGLM系列来自清华大学KEG实验室与智谱AI，采用不同于GPT自回归的自回归填空预训练目标81% similarUnverifiedGLM基于自回归空白填充预训练，融合了BERT的双向理解能力与GPT的生成能力，通过二维位置编码区分原始序列位置与空白内部位置77% similarVerifiedBERT's bidirectional encoding capability stems from the masked language model (MLM) objective used during its pre-training phase, which contrasts with the unidirectional autoregressive pre-training of the GPT series.77% similarUnverifiedGPT 系列、Llama 等模型的核心能力绝大部分来自预训练而非后续微调阶段75% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/22146API
curl https://kongchang.com/api/v1/knowledge/claims/22146MCP
get_claim(id=22146)