Unverified50% confidenceFactExact time
The development of model distillation is divided into three stages: the first stage (2015–2018) focused on model compression applied to computer vision, the second stage (2019–2021) expanded into the NLP domain, and the third stage (2022 to present) entered the large language model era and evolved into capability transfer
1
Sources
50%
Confidence
Long-term
Relevance
9/10/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe rise of Prompt Engineering is closely tied to the explosion of large language models (LLMs) in 2022–2023.68% similarUnverified从2022年的Stable Diffusion到2023年的ChatGPT再到2024年的多模态模型，AI领域几乎每隔几个月就有颠覆性的新工具和框架出现67% similarUnverified离散扩散模型发展经历D3PM（2021年Austin等人）、MDLM（2023年Sahoo等人）、SEDD（2023年Lou等人）等里程碑，但模型规模通常不超过1B参数65% similarUnverified多Agent范式在2023至2024年间随着AutoGPT、CrewAI、LangGraph等开源框架的兴起而快速走向成熟65% similarUnverified结构化提示词（Structured Prompting）是2023年初大模型应用早期流行的一种技巧65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/887107API
curl https://kongchang.com/api/v1/knowledge/claims/887107MCP
get_claim(id=887107)