Text this: On the Validity of Traditional Vulnerability Scoring Systems for Adversarial Attacks Against LLMs