Enjinx
  • Enjinx
    • What is Enjinx ?
    • Setup and Installation
      • Install on VSCode
      • Install on JetBrains IDEs
      • Subscription Plans
      • Sign in
      • Extension Settings
      • Uninstall Enjinx
    • Enjinx Chat
      • Focus
        • Current File Focus
        • Git Diff Focus
      • Context
        • Add Context
        • Add image as Context
      • Commands
        • /ask
        • /changelog
        • /commit
        • /describe
        • /docstring
        • /enhance
        • /explain
        • /find-on-github
        • /generate-best-practices
        • /help
        • /improve
        • /issues
        • /quick-test
        • /recap
        • /review
        • /test-suite
      • Chat History
      • Model Selection
      • Company Codebase
    • Code Completion
    • Coding Agent
      • Tasks
      • Continuous Code Improvement
    • Test Generation
      • Using Test Generation
      • Behaviors
      • Test Suite
      • Configuration
      • Example Test
      • Context
      • Running Tests
    • Data Sharing
  • Enjinx cover
    • What is Enjinx Cover ?
    • Setup
      • GitHub Action
      • CLI
    • Feature Flags
    • Database Usage
    • Coverage Report
    • Examples
  • Alpha Codium
    • What is Alpha Codium?
    • Setup
    • Usage
      • Configuration
      • Solving Problems
      • Solving the entire dataset
      • Evaluation
    • Technical Q&A
    • Research Paper
    • Examples
Powered by GitBook
On this page
  1. Alpha Codium

Technical Q&A

Prompt vs. Flow Engineering

Q: How much time was spent on "prompt engineering" compared to "flow engineering"?

A: Approximately 95% of the effort was dedicated to high-level design and flow engineering.

Data Leakage Prevention

Q: How is data leakage avoided?

A: The test set includes problems from after September 2021, aligning with GPT-4's data cutoff, ensuring no leakage specifically for GPT-4.

Language Agnosticism

Q: Is AlphaCodium limited to specific programming languages?

A: No, the flow is language agnostic and can be applied to any language.

Context Window Management

Q: How is the context window managed?

A: Models with an 8192 token context window were used. Larger contexts may lead to some information being ignored.

Iteration on Code vs. Tests

Q: Why iterate only on generated code and not on the AI-generated tests?

A: For code problems in CodeContests, iterating on code is more beneficial than on tests, which are simple input-output pairs.

PreviousEvaluationNextResearch Paper

Last updated 5 months ago