claim
The GraphEval framework categorizes an entire LLM output as containing a hallucination if at least one triple within the constructed Knowledge Graph is flagged as inconsistent with the provided context.

Authors

Sources

Referenced by nodes (5)