Prepare and test your skills
Prepare and test your skills
Free sample question, answered and explained below. The Generative AI Leader practice bank has 2,191 more like it. Try 10 free: instant score, no account.
An insurance company is preparing a specialized dataset to tune a generative AI assistant for its customer support team. During testing, the assistant frequently produces inaccurate claim summaries, invents policy rules, and provides inconsistent guidance for regional policies that contained outdated or missing information in the prepared dataset.
Which factor directly causes these invalid and suboptimal outputs from the generative model?
Data quality and tuning integrity refers to the accuracy, completeness, relevance, and consistency of the dataset used to train, tune, or ground a machine learning model. In generative artificial intelligence, models learn patterns, factual associations, and contextual rules directly from their underlying training and tuning datasets.
Generative AI models strictly reflect the data they are tuned on. If input datasets contain errors, gaps, or outdated policies, the model will inevitably produce invalid outputs regardless of prompt formatting or underlying infrastructure.
Feature serving pipelines are operational data engineering workflows used in traditional machine learning to retrieve and export feature values for batch processing or real-time online inference.
Input token limits define the maximum context length that a large language model can process in a single request. Every model has an architectural boundary for how many words, characters, or subwords it can accept in a prompt.
Few-shot prompting is a prompt engineering strategy where a small number of concrete examples are provided in the prompt to guide the model's formatting, tone, and response structure.
Answer them yourself and get your score the moment you finish. No account needed.
Full access includes
Keep the momentum going with these hand-picked practice scenarios
Last updated