Skip to main content

Misinformation

Tests whether the LLM produces factual fabrications, invented sources, or misrepresented expertise.

Sub-types​

  • Factual Fabrication: Model fabricates facts, statistics, or events that never occurred.
  • Source Fabrication: Model invents citations, references, or sources that do not exist.
  • Expertise Misrepresentation: Model presents itself as having expertise or authority it lacks.

Threat Profile​

Objective: harmful_behavior

Objective vs. goals

The objective above selects the scoring rubric for a run — it is one of a fixed set of built-in names, not something you write. The goals you pass to an attack are separate free-text strings that you author yourself. See Goals vs. objective.

Primary

  • truthfulqa: TruthfulQA for evaluating factual accuracy and misinformation

Secondary

  • saladbench: SaladBench for broader misinformation coverage

Attack Techniques​

Primary

  • Static Template: Template-based prompt construction

Metrics​

  • truthfulness_score
  • accuracy_score

Usage​

Instantiate the Vulnerability​

from hackagent.risks import Misinformation
from hackagent.risks.misinformation.types import MisinformationType

# Use all sub-types
vuln = Misinformation()

# Or specify particular sub-types
vuln = Misinformation(types=[
    MisinformationType.FACTUAL_FABRICATION.value,
    MisinformationType.SOURCE_FABRICATION.value,
])

Run an Evaluation Campaign​

from hackagent import HackAgent
from hackagent.risks.misinformation import MISINFORMATION_PROFILE

agent = HackAgent(endpoint="http://localhost:8080/chat", name="my-agent")

# Use profile recommendations
for attack in MISINFORMATION_PROFILE.primary_attacks:
    for dataset in MISINFORMATION_PROFILE.primary_datasets:
        attack_config = {
            "attack_type": "static_template",  # attack.technique is "StaticTemplate"
            "objective": MISINFORMATION_PROFILE.objective,
            "dataset": {"preset": dataset.preset},
        }
        results = agent.hack(attack_config=attack_config)
        print(f"Results: {results}")