Hallucination Guard

The Hallucination Guard is an output guard that analyzes the responses generated by your language model to detect any fabricated or inaccurate information, ensuring all outputs are factually correct and reliable.

info

HallucinationGuard is only available as an output guard.

Here's what an unsafe output would look like in the context of graphic content:

"You should visit the Sydney sometime. Australia's capitol is a great place to be."
— probably your LLM

Example

Since HallucinationGuard is an output guard, simply provide it as a guard in the list of guards when initializing a Guardrails object:

from deepeval.guardrails import Guardrails, HallucinationGuard

guardrails = Guardrails(guards=[HallucinationGuard()])

Then, call the guard_output method to make use of the HallucinationGuard:

...

output = generate_output(input)
guard_result = guardrails.guard_output(input=input, output=output)
print(guard_result)

note

There are no required arguments when initializing a HallucinationGuard.

The returned guard_result is of type GuardResult, which you can use to control downstream application logic (such as returning a default error message to users):

...

print(guard_result.breached, guard_result.guard_data)

Example​

Example