Package com.tnsai.evaluation.evaluators.safety

Class HallucinationEvaluator

java.lang.Object

com.tnsai.evaluation.evaluators.AbstractJudgeEvaluator

com.tnsai.evaluation.evaluators.safety.HallucinationEvaluator

All Implemented Interfaces:

Evaluator

public final class HallucinationEvaluator
extends AbstractJudgeEvaluator

Detects hallucinated content in agent responses by checking factual claims against provided context.

Unlike FaithfulnessEvaluator, this evaluator works with any context source (not just RAG retrieval) and focuses specifically on safety — flagging fabricated facts, invented references, or claims contradicting the context.

Context is provided via metadata.get("context") as a String or List<String>. If no context is provided, evaluates against the agent's own internal consistency.

Score is inverted: 1.0 = no hallucination, 0.0 = fully hallucinated.

Since:

0.3.0

Nested Class Summary

Nested classes/interfaces inherited from interface com.tnsai.evaluation.spi.Evaluator

Evaluator.EvaluationInput

Constructor Summary

Constructors

Constructor	Description
HallucinationEvaluator(LLMJudge judge)

Method Summary

Methods inherited from class com.tnsai.evaluation.evaluators.AbstractJudgeEvaluator

evaluate, name

Methods inherited from class java.lang.Object

equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Details

HallucinationEvaluator

public HallucinationEvaluator(LLMJudge judge)