Unverified50% confidenceFactExact time
GLM系列由清华大学与智谱AI联合研发，其架构核心是对标准Transformer的自回归空白填充预训练目标的改造
1
Sources
50%
Confidence
Long-term
Relevance
7/8/2026
First Seen
Sources
AMD MI355X运行GLM5.2：推理成本仅为英伟达Blackwell的一半
hackernewshackernews7/3/2026
Related Claims
UnverifiedGLM 5.2 was jointly developed by Tsinghua University and Zhipu AI, with specialized optimization for code generation and tool-calling scenarios.75% similarVerifiedThe GLM series is a large language model jointly developed by Tsinghua University and Zhipu AI, built on the Autoregressive Blank Infilling training objective.75% similarUnverifiedRWKV将RNN的线性推理效率与Transformer的并行训练能力融合70% similarUnverified大模型核心原理包括Transformer架构、预训练、SFT监督微调、RLHF人类反馈强化学习等关键概念70% similarUnverifiedLeaps in model capability stem from Transformer architecture optimization, improvements in training data scale and quality, and the maturation of post-training alignment techniques such as RLHF and DPO.70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/219932API
curl https://kongchang.com/api/v1/knowledge/claims/219932MCP
get_claim(id=219932)