Quiz on Large Language Models (LLM)

What is the primary capability of a Large Language Model (LLM)? Choose the correct answer.

• It performs hardware-level automation tasks.

• It only works with numerical data.

• It specializes in understanding and generating human language.

• It replaces all traditional machine learning models.

What best describes SAP’s approach to Large Language Models (LLMs) within SAP AI Core? Choose the correct answer.

• SAP limits the use of LLMs to preconfigured chatbot scenarios only.

• SAP provides LLMs as part of a SaaS offering, enabling integration into business applications without needing to design or train models from scratch.

• SAP’s LLMs require businesses to host and manage the models on their own infrastructure.

• SAP’s LLMs cannot be combined with other AI models for complex business tasks.

Which of the following is a risk associated with using Large Language Models (LLMs) in SAP Business AI? Choose the correct answer.

• Improved customer experience through personalized assistance.

• Ability to analyze large amounts of data quickly.

• Potential for data privacy concerns when handling sensitive information.

• Enhanced scalability and efficiency in business processes.

Which of the following are key use cases of Large Language Models (LLMs) in SAP AI Core? Choose the correct answer.

• Managing network infrastructure and hardware provisioning.

• Drafting emails, writing code, translating languages, and generating context-aware content.

• Monitoring system uptime and hardware diagnostics.

• Performing real-time data backups and system migrations.

Which of the following best describes SAP’s product development process for LLM scenarios? Choose the correct answer.

• A three-step approach that skips customer validation to accelerate deployment.

• An iterative process focusing only on ideation and experimentation of generative AI concepts.

• A structured four-phase process including Ideation, Validation, Realization, and Continuous Improvement.

What is the recommended approach when applying prompt engineering and optimization techniques to improve LLM performance for your use case? Choose the correct answer.

• Use a single optimization technique and avoid iterative evaluation to reduce complexity.

• Follow a strict linear process with fixed steps from prompt engineering to fine-tuning.

• Apply a combination of techniques such as prompt engineering, RAG, and fine-tuning in an iterative way based on ongoing evaluations.

• Rely entirely on pre-trained models and avoid modifying prompts or training data.

Which of the following best describes the difference between Chain-of-Thought (CoT) and Tree-of-Thought (ToT) techniques in prompt engineering? Choose the correct answer.

• CoT breaks a problem into smaller prompts for linear reasoning, while ToT creates a branching structure to explore multiple reasoning paths in parallel.

• CoT relies on random prompts, while ToT requires predefined questions only.

• Both CoT and ToT are used only for simple factual queries with no reasoning steps.

• CoT and ToT are techniques used only for training the LLM, not for crafting prompts.