IBM watsonx School Quiz

What are the 3 types of prompt development modes that are supported by IBM watsonx.ai?

• Structured, Freeform, Chat

• Structured, Few-Shot, Chat

• Structured, One-Shot, Multi-Shot

• Naive, One-Shot, Multi-Shot

What are the Large Language Model families offered via IBM watsonx.ai?

• ChatGPT, Copilot, Gemini, DeepSeek

• Granite, Llama, Flan, Mistral

• LlamaIndex, LongChain, InstructLab, CrewAI

• IBM watsonx.ai, watsonx.data, watsonx.governance, watsonx Orchestrator

Which of the following model categories do not belong to the Granite Family?

• Vision Models

• Guardrail Models

• Geospatial Models

• Mixture of Dummy Models

What are the most informative KPIs for evaluating the quality of Large Language Models?

• The Model Card metadata of the LLM

• ROUGE, BLEU, Precision, Recall, F1 Score

• #Input Tokens, #Output Tokens, LLM Call Time

• ROUGE, BLEU, BERTScore

Which expression is correct?

• When using Prompt Tuning in IBM watsonx.ai, the model weights are changed.

• The Model Transparency Index evaluates the 10 most popular LLM families.

• InstructLab provides taxonomy-based data curation and utilizes synthetic

  data generation to tune a model.

• Agentic AI provides an essential building block for the appropriate application of LLMs.