hallucination-check
LLM hallucination detector with dual strategy (UQLM + rule-based fallback). Scores any AI output's confidence and flags potential hallucination risks.
