Pre-LLM Validation

Pre-LLM validation checks user input for semantic coherence by verifying term connectivity within a local knowledge graph, ensuring prompts are well-formed before the LLM call. This prevents ill-formed inputs from wasting tokens or causing hallucinations.

Hallucinations and off-topic responses often occur when inputs lack domain-term coverage. Pre-flight validation assesses user queries for domain-specific terms and semantic connectivity before the LLM call, ensuring only coherent, relevant prompts are processed.

Domain-term coverage is checked by applying pre-flight validation to assess the presence of domain-specific terms and concepts. The validation implements domain-term matching to confirm the user query aligns with the intended domain before generating a response.

Yes, pre-LLM validation optionally suggests alternatives or refinements for unclear prompts. It provides guided corrections by using the terrraphim-agent validate tools to check connectivity and domain terms, refining user queries before they reach the LLM.

Pre-flight validation is not necessary for open-ended, creative, or general conversational prompts where strict domain alignment is not required. It is designed for domain-specific queries where semantic coherence and domain-term coverage are critical to prevent token waste.