Log på
indsigt - Hallucination Detection Framework
Zero-Shot Multi-task Hallucination Detection Study at SemEval-2024
Detecting and defining hallucinations in NLG tasks using a quantitative framework.
Unsupervised Real-Time Hallucination Detection in Large Language Models
The author introduces MIND, an unsupervised training framework utilizing internal states of Large Language Models for real-time hallucination detection. The approach outperforms existing methods in detecting hallucinations.
Produkter
© 2024 by Linnk AI