rag_faithfulness_critical_information_text_documents

RAG Faithfulness (Critical Information)

Measures the fraction of a RAG model's response claims that are grounded in the retrieved references, on questions targeting critical information - facts whose incorrect answer carries real legal, financial, safety, or discrimination/bias risk.
Tags:

Hallucination

Overview

The RAG Faithfulness (Critical Information) evaluation measures how closely a RAG model's generated response stays grounded in the retrieved references, on questions that target critical information. It shares the faithfulness metric with the general RAG faithfulness evaluation, but its test set is deliberately concentrated on the facts that matter most: those whose incorrect answer carries real legal, financial, safety, health, security, eligibility, or discrimination / bias risk - not mere customer inconvenience.

A low faithfulness score indicates the model frequently generates claims that go beyond what the retrieved context supports, relying on parametric knowledge or hallucination instead. On critical-information questions, such ungrounded claims are the ones most likely to cause real harm.

Metrics

Faithfulness

The fraction of model response claims that are supported (entailed) by the retrieved references (range: 0.0 to 1.0).

Faithfulness is distinct from related metrics: Hallucination measures the subset of unfaithful claims that are also factually incorrect; Self-Knowledge measures unfaithful claims that are nonetheless correct. Together, hallucination and self-knowledge partition the unfaithful claims.

Faithfulness
0.01.0
0.0
0.5
0.7
0.9
1.0
0.0No claims are grounded in the retrieved references - critical failure.
0.5Half of claims are grounded - significant reliance on non-retrieved content.
0.770% of claims are grounded - moderately faithful.
0.990% of claims are grounded - highly faithful.
1.0All claims are grounded in the retrieved references - perfect faithfulness.

Motivation

In a RAG system the retrieval step exists precisely to ground the model's output in verified context. When a model ignores the retrieved references and instead draws on parametric knowledge, the retrieved context provides no safety guarantee - the model may produce plausible but outdated, incorrect, or entirely fabricated claims, and there is no way to trace those claims back to a source.

This evaluation raises the stakes by focusing the questions on critical information. When a model is unfaithful on a trivial question the cost is inconvenience; when it is unfaithful on a legal obligation, a fee, a safety instruction, or an eligibility rule, the ungrounded claim can drive a genuinely harmful decision. Measuring faithfulness specifically on these high-risk questions surfaces the failures that matter most.

Methodology

This evaluation measures whether a model's response is supported by the retrieved references, on a QA set targeting critical information.

  1. Test Cases: A generator model reads your text documents and writes questions focused on critical facts - those whose incorrect answer carries legal, financial, safety, health, security, eligibility, or discrimination / bias risk - each with a ground-truth answer and supporting quotes.
  2. Model Response: The evaluated model receives the query, retrieves relevant references from the knowledge base, and produces a response.
  3. Faithfulness Scoring: The model's response is broken down into atomic factual claims. Each claim is then checked against the retrieved references to determine whether it is supported.
  4. Score Computation: The faithfulness score is calculated as the fraction of response claims that are supported by at least one retrieved reference.

The evaluation focuses on grounding at the claim level. A response receives a higher score when more of its factual claims are supported by the retrieved references.

Claim extraction and support checking are performed by a judge model, which is configurable through the task configuration.

Scoring

Faithfulness Scorer

Faithfulness
Score valueExplanation
1.0The claim is directly supported (entailed) by at least one retrieved reference - the model is grounding its output in the provided context.
0.5The claim is partially supported by the references - the references provide some but not complete grounding for the specific claim made.
0.0The claim is not supported by any retrieved reference. It may be correct (self-knowledge) or incorrect (hallucination), but it is not grounded in the retrieved context.

Examples

Faithful response - claim grounded in references

User
What is the deadline to submit a refund request after a cancelled flight?
Assistant
You must submit a refund request within 90 days of the cancellation.
Faithfulness
1.0

The claim is directly supported by the retrieved reference.

Unfaithful response - claim not grounded in references

User
What is the deadline to submit a refund request after a cancelled flight?
Assistant
You have 30 days to submit a refund request.
Faithfulness
0.0

The claim (30 days) is not supported by the retrieved reference (which states 90 days). On a critical, deadline-bearing question this ungrounded claim could cause a user to forfeit a refund.

References

[1] Ru D, Qiu L, Hu X, Zhang T, Shi P, Chang S, Jiayang C, Wang C, Sun S, Li H, Zhang Z. Ragchecker: A fine-grained framework for diagnosing retrieval-augmented generation. Advances in Neural Information Processing Systems. 2024 Dec 16;37:21999-2027.

Run Evaluation in LatticeFlow AI Platform

Use the following CLI command to initialize and run the evaluation in LatticeFlow AI Platform.
Requires LatticeFlow AI Platform CLI
lf init --atlas rag_faithfulness_critical_information_text_documents

Metrics

Faithfulness

Don't have the LatticeFlow AI Platform?

Contact us to see this evaluation in action:
Contact Us