IntensiveCoLearning / IntensiveCoLearning/AI-Web3-School

Daily check-in: 2026-05-21

Open
#30 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
8
Forks
40
PR merge metrics
No merged PRs in 30d

Description

Day 1 打卡 — LLM + Prompt

学了什么
  • LLM 是概率模型,不是知识数据库——它每一步只选"下一个最顺眼的词",不是"找正确答案"
  • 幻觉(Hallucination)是架构决定的必然行为——模型没有"我不知道"的选项
  • Prompt 是接口设计,不是魔法咒语。Instruction 四段式:任务目标 / 可用输入 / 禁止行为 / 输出格式
  • Temperature 对比实验:T=0 确定 vs T=0.7 平衡 vs T=1.5 随机
  • 结构化输出(JSON/schema)让机器可检查,但安全边界靠代码层
核心认知链

LLM 是概率模型 → 它不知道自己在说什么,只是选下一个最顺眼的词 → 所以它天然会编造(hallucination) → 所以它需要外部状态管理(数据库/链上) → 但 AI 自己判断不了"什么该存、什么该变" → 这个判断(决断)只能由人来做 → 所以 LLM 是推理层,人类是决策层

关键思考
  1. 小说崩盘 = LLM 幻觉——都是"只关注当前这一步怎么顺,不关注之前承诺过什么"
  2. 小说本质是变化过程不是静态事实——"变化轨迹管理"是无人解决的工程问题
  3. Web3 保证不可篡改已实现,但"什么该存"的判断仍是人的活
  4. Prompt 的软边界可被后续输入不断改变概率——LLM 输出是概率驱动的

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

This issue contains a Day 1 learning log about LLMs, prompts, hallucinations, and Web3, but it names no files, tests, or implementation entry points. There is no requested change or definition of done, so a contributor should first confirm whether this is intended as project content and what update is expected.

Written by the indexing model from the issue text.

Assessment

Tech stack
blockchain, machine-learning
Domain
ai, blockchain, content, documentation
Issue type
Documentation
Difficulty
1/5
Estimated time
Under an hour
Activity status
Quiet
Clarity
Needs clarification
Newbie friendliness
20/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.