intrinsics

Intrinsics Examples

This directory contains examples for using Mellea's intrinsic functions - specialized model capabilities accessed through adapters.

Files

intrinsics.py

Core example showing how to directly use intrinsics with adapters.

Key Features:

• Creating and adding adapters to backends
• Using Intrinsic component for specialized tasks
• Working with Granite Common adapters (aLoRA-based)
• Understanding adapter output formats

answerability.py

Checks if a question can be answered given the context.

citations.py

Validates and extracts citations from generated text.

context_relevance.py

Assesses if retrieved context is relevant to a query.

hallucination_detection.py

Detects when model outputs contain hallucinated information.

query_rewrite.py

Rewrites queries for better retrieval or understanding.

uncertainty.py

Estimates the model's certainty about answering a question.

requirement_check.py

Detect if text adheres to provided requirements.

policy_guardrails.py

Checks if a scenario is compliant/non-compliant/ambiguous with respect to a given policy,

guardian_core.py

Uses the guardian-core LoRA adapter for safety risk detection, including prompt-level harm, response-level social bias, RAG groundedness, and custom criteria.

factuality_detection.py

Detects if the the model's output is factually incorrect relative to context.

factuality_correction.py

Corrects a factually incorrect response relative to context.

context_attribution.py

Identifies sentences in conversation history and documents that most influenced the response.

Concepts Demonstrated

• Intrinsic Functions: Specialized model capabilities beyond text generation
• Adapter System: Using LoRA/aLoRA adapters for specific tasks
• RAG Evaluation: Assessing retrieval-augmented generation quality
• Quality Metrics: Measuring relevance, groundedness, and accuracy
• Backend Integration: Adding adapters to different backend types

Basic Usage

from mellea.backends.huggingface import LocalHFBackend
from mellea.backends.adapters.adapter import IntrinsicAdapter
from mellea.stdlib.components import Intrinsic
import mellea.stdlib.functional as mfuncs

# Create backend and adapter
backend = LocalHFBackend(model_id="ibm-granite/granite-4.0-micro")
adapter = IntrinsicAdapter("requirement_check",
                               base_model_name=backend.base_model_name)
backend.add_adapter(adapter)

# Use intrinsic
out, new_ctx = mfuncs.act(
    Intrinsic("requirement_check",
              intrinsic_kwargs={"requirement": "The assistant is helpful."}),
    ctx,
    backend
)

Available Intrinsics

• requirement_check: Validate requirements (used by ALoraRequirement)
• answerability: Determine if question is answerable
• citations: Extract and validate citations
• context_relevance: Assess context-query relevance
• hallucination_detection: Detect hallucinated content
• query_rewrite: Improve query formulation
• uncertainty: Estimate certainty about answering a question
• policy_guardrails: Determine if scenario complies with policy
• guardian-core: Safety risk detection (harm, bias, groundedness, custom criteria)
• factuality_detection: Detect factually incorrect responses
• factuality_correction: Correct factually incorrect responses
• context-attribution: Identify context sentences that most influenced response

Related Documentation

• See mellea/stdlib/components/intrinsic/ for intrinsic implementations
• See mellea/backends/adapters/ for adapter system
• See docs/dev/intrinsics_and_adapters.md for architecture details