top of page
TruthChecker
Detect, Explain and Correct LLM hallucinations
Play Video
Strongest available approach and solution in the industry
Model based hallucination detection
Automatic dataset generation and evaluation
Hands-off active learning
Use with API based LLMs (GPT4, etc) and Open Source LLMs (Mistral, Llama, Zephyr, etc)
Use with knowledge bots, summarization, etc
3 simple ways to use
TruthChecker
Quick test using API
Try using a test API with our base Hallucination Manager
​
(available via NVIDIA NeMo Guardrails soon)
High accuracy setup to use with your RAG pipeline
Build you custom hallucination detection model for your RAG pipeline
Highest accuracy, custom setup for enterprise use cases
Build fully tuned, GPT-4 level accuracy system for your RAG pipeline, summarization or other tasks
bottom of page