What specific behaviors does your framework target? (Be precise: not "disinformation" but specific conducts or mechanisms.)