evals.span_templates

HALLUCINATION_SPAN_PROMPT_TEMPLATE

Prompt template designed to evaluate language model (LLM) hallucination and intended to be used with llm_classify() and run_evaluations().

The template helps in structuring input to assess the accuracy of information generated by the model, particularly focusing on detecting instances where the model generates false or misleading information (hallucinations).

QA_SPAN_PROMPT_TEMPLATE

Prompt template designed to evaluate the performance of language models (LLMs) on question answering (QA) tasks and intended to be used with llm_classify() and run_evaluations().

The template helps assess how effectively the LLM can understand/respond to questions and provide accurate and relevant answers.