Verified80% confidenceFactExact time
Fine-tuning refers to performing secondary training on a foundation model that has completed general-text pre-training, using domain-specific labeled data to adjust the model's parameters
8
Sources
80%
Confidence
Long-term
Relevance
7/19/2026
First Seen
Sources
Codex实战指南：AI编程助手从入门到进阶全攻略
Related Claims
Unverified微调是指在已具备通用语言理解能力的基础模型之上，用特定领域的专业数据进行二次训练，使模型在保留通用能力的同时对特定任务表现出更高的专业精度80% similarVerified大语言模型的预训练阶段本质上属于自监督学习，是一种特殊的非监督学习，模型通过预测下一个词来从海量文本中自动学习语言规律77% similarUnverifiedFoundation models refer to models pretrained on large-scale general-purpose corpora that can be adapted to a variety of downstream tasks through fine-tuning or prompting, and are usually categorized by modality into text, image, audio and other types.76% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/559013API
curl https://kongchang.com/api/v1/knowledge/claims/559013MCP
get_claim(id=559013)