Skip to content
Logo LogoRagas
⌘ K
Logo LogoRagas
  • 🚀 Get Started
    • Installation
    • Generate a Synthetic Test Set
    • Evaluating Using Your Test Set
    • Monitor Your RAG in Production
  • 📚 Core Concepts
    • Metrics-Driven Development
    • Metrics
      • Faithfulness
      • Answer Relevance
      • Context Precision
      • Context utilization
      • Context Recall
      • Context entities recall
      • Noise Sensitivity
      • Answer semantic similarity
      • Answer Correctness
      • Aspect Critique
      • Domain Specific Evaluation
      • Summarization Score
    • Prompt Objects
    • Automatic prompt Adaptation
    • Synthetic Test Data generation
    • Utilizing User Feedback
  • 🛠️ How-to Guides
    • Customizations
      • Bring Your Own LLMs and Embeddings
      • Using Ragas Critic Model instead of GPT-4
      • Max Workers, Timeouts, Retries and more with RunConfig
      • Using Azure OpenAI
      • Using Amazon Bedrock
      • Vertex AI
    • Applications
      • Building HF Dataset with your own Data
      • Understand Cost and Usage of Operations
      • Compare Embeddings for retriever
      • Compare LLMs using Ragas Evaluations
      • Write custom prompts with ragas
      • Automatic language adaptation
      • Explainability through Logging and tracing
      • Adding to your CI pipeline with Pytest
    • Integrations
      • LlamaIndex
      • Langchain
      • Langsmith
      • Phoenix (Arize)
      • Langfuse
      • Athina AI
      • Zeno
      • Tonic Validate
      • Haystack
      • OpenLayer
      • Helicone
  • 📖 References
    • Evaluation
    • Metrics
    • RunConfig
    • Integrations
  • ❤️ Community
Ragas
/
📚 Core Concepts
/
Metrics

Metrics¶

Component-Wise Evaluation¶

Just like in any machine learning system, the performance of individual components within the LLM and RAG pipeline has a significant impact on the overall experience. Ragas offers metrics tailored for evaluating each component of your RAG pipeline in isolation.

evol-generate

  • Faithfulness

  • Answer relevancy

  • Context recall

  • Context precision

  • Context utilization

  • Context entity recall

  • Noise Sensitivity

  • Summarization Score

  • Faithfulness
    • Example
    • Calculation
    • Faithfullness with HHEM-2.1-Open
  • Answer Relevance
    • Example
    • Calculation
  • Context Precision
    • Example
    • Calculation
  • Context utilization
    • Example
    • Calculation
  • Context Recall
    • Example
    • Calculation
  • Context entities recall
    • Example
    • Calculation
  • Noise Sensitivity
    • Example
    • Calculation
  • Answer semantic similarity
    • Example
    • Calculation
  • Answer Correctness
    • Example
    • Calculation
  • Aspect Critique
    • Example
    • Calculation
  • Domain Specific Evaluation
    • Example
  • Summarization Score
    • Example
Metrics-Driven Development
Faithfulness

On this page

  • Component-Wise Evaluation

© 2023, ExplodingGradients Built with Sphinx 6.2.1