fine-tuning-decisions
Skill0 starsChanged 3 months agoMITView on GitHub
Decide whether to fine-tune vs prompt engineering vs RAG, and how to plan a fine-tune. 判断该不该微调、怎么规划微调。
In this repository14 other files it gives its agents.
- ai-code-reviewSkill
- ai-cost-estimatorSkill
- ai-news-digestSkill
- ai-paper-highlightsSkill
- ai-tools-radarSkill
- ai-writing-assistantSkill
- embedding-searchSkill
- llm-evaluationSkill
- llm-security-checklistSkill
- mcp-server-guideSkill
- model-comparisonSkill
- prompt-engineering-playbookSkill
- rag-pipelineSkill
- structured-output-validationSkill
The file
Files
Fine-tuning Decisions | 该不该微调
Decide whether fine-tuning is worth it for a problem — or whether prompt engineering / RAG is the smarter move — and how to plan one.
判断一个问题该微调、还是该调 prompt / 上 RAG,以及真要微调该怎么规划。
When to use | 适用场景
- 用户问「要不要微调一个模型」
- 现有 prompt/RAG 方案效果到瓶颈
Workflow | 工作流程
- 先诊断失败模式:是格式不对 / 知识缺失 / 风格不对 / 能力不够?
- 对照选择:
- 知识缺失 → RAG(微调学不进去事实)
- 格式/风格/输出结构 → 提示工程 或 微调
- 需要新能力(新语言、新领域格式)→ 微调
- 若微调:评估数据量(通常 ≥ 500 条高质量样本)、算力/成本、评测集。
- 给出建议 + 最小可行验证方案。
Tips | 提示
- 微调改的是「行为」,不是「知识」,知识问题先找 RAG。
- 先小数据跑通再扩,别一上来就买大算力。
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

