Skip to content

grimwiz/ai-training

Repository files navigation

LLM Risk Guide for Home Users

A small static Next.js site that teaches common LLM failure modes in everyday home use.

What is included

  • Home page (/) with intro and navigation.
  • Risks index (/risks) with search/filter and a home-user-first ordering.
  • Detail pages (/risks/[id]) with plain-language explanation, checklist, sources, and demo.
  • About page (/about) with usage steps and source information.
  • Scripted, no-API demos for the top 6 home risks in demos/.

Tech stack

  • Next.js (App Router) + TypeScript
  • Tailwind CSS
  • PapaParse for CSV parsing at build/runtime on the server side
  • No backend, no database, no external LLM calls

Run locally

npm install
npm run dev

Then open http://localhost:3000.

Build

npm run build

Available scripts

  • npm run dev: Start local development server.
  • npm run build: Build production output.
  • npm run start: Start production server from a built app.
  • npm run lint: Run Next.js lint command.

Update llm_risk_taxonomy.csv

  1. Replace or edit llm_risk_taxonomy.csv in the repo root.
  2. Keep these headers intact:
    • ID, Title, What_happens, Why_it_happens, Training_vignette, Test_prompt, Mitigation_checklist, Source_URLs
  3. Run npm run dev or npm run build to verify parsing still works.

Add a new demo in demos/

  1. Create demos/<risk-id>.ts that exports a DemoDefinition.
  2. Include:
    • headline
    • setup
    • input
    • controls
    • outputPanels
    • teachingPoints
  3. Register it in demos/index.ts by mapping the risk ID (for example R21) to your export.
  4. Open /risks/<ID> and verify the interactive demo appears.

Project structure

  • app/: Routes and pages.
  • components/: Reusable UI (RiskCard, RiskDemo, index controls).
  • lib/risks.ts: CSV loading, ordering, and risk view model shaping.
  • demos/: Scripted interactive demo definitions by risk ID.
  • llm_risk_taxonomy.csv: Source data.

About

website for training users how to identify and avoid ai problems (LLM)

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors