QA Relevance LLM

Features

1 min read

QA Relevance LLM

The QA Relevance LLM validator compares the similarity between the input prompt and LLM-generated output. After comparison, the validator generates a confidence score representing the contextual relevance of the response. This validator ensures that LLM outputs are relevant to user queries, fostering a good user experience. The validator measures relevance in three steps:

1. The scanner converts both the prompt and the output into vector embeddings.

2. It calculates the cosine similarity between these embeddings.

3. This similarity score is compared against a predefined threshold to determine relevance.



QA Relevance LLM

© 2024 Wisecube AI

© 2024 Wisecube AI

© 2024 Wisecube AI