A. To limit the maximum text length that the model generates by ensuring concise responses.
B. To determine the number of tokens the model can process at once by influencing the complexity
and length of inputs and outputs.
C. To filter out potentially harmful or inappropriate content from the model’s output based on the
desired level of filtering.
D. To control the creativity and randomness of the model’s output by adjusting the diversity of word
choices.
C
A. Collect a larger and more diverse dataset for the gen Al model.
B. Fine-tune the gen Al model.
C. Implement explainable gen Al policies.
D. Develop fairness assessments for the gen Al model.
C
A. LLMs excel in highly specific technical tasks requiring deep, singular domain expertise.
B. LLMs learn and generalize effectively from small datasets for niche applications.
C. LLMs have strong inherent logical reasoning and problem-solving abilities without extra
prompting.
D. LLMs are trained on vast datasets, enabling broad language and context understanding, and
adaptability across many tasks.
D
A. Fine-tuning
B. Prompt engineering
C. Retrieval-augmented generation (RAG)
D. Human-in-the-loop (HITL)
C
A. Gemini
B. Gemma
C. Veo
D. Imagen
D
A. Supervised learning
B. Deep learning
C. Unsupervised learning
D. Reinforcement learning
D
A. To provide the computing power for training and running advanced Al models.
B. To be the user interface for interacting with Al models.
C. To be a smart system that can analyze, use tools, and make decisions to reach goals.
D. To be a central storage place for the data that Al models use.
C
A. Google Cloud’s emphasis on an open approach within its Al offerings.
B. Google Cloud’s commitment to tightly integrated, proprietary Al solutions.
C. Google Cloud’s strategy prioritizing fully managed Al services that simplify the user experience.
D. Google Cloud’s primary focus on automating Al workflows.
A
A. NotebookLM
B. Gemini app
C. Vertex Al Search
D. Gemini for Google Workspace
A
A. Learning from labeled data with correct output pairs.
B. Learning by identifying patterns in unlabeled data.
C. Learning through interaction and feedback.
D. Learning by training on vast data to generate new content.
C
A. Gemini for Google Workspace
B. Google Agentspace
C. Vertex Al Search
D. Conversational Agents
C
A. Cloud Functions
B. Vertex Al
C. Google Agentspace
D. BigQuery
B
A. Gemma
B. Gemini
C. Imagen
D. Veo
D
A. Bias
B. Knowledge cutoff
C. Data dependency
D. Hallucinations
D
A. A physical device that houses the hardware components of a gen Al system.
B. A complex algorithm trained on vast amounts of data to learn patterns and relationships.
C. A user interface that allows users to interact with a gen Al system.
D. A set of rules and guidelines governing responsible development and use of gen Al.
B
A. Vertex Al Platform
B. Google Cloud Contact Center as a Service
C. Conversational Al
D. Vertex Al Search
B
A. Agentspace primarily focuses on enhancing external customer engagement through Al-powered
chatbots.
B. Agentspace directly manages the underlying infrastructure and hardware required for Al model
training.
C. Agentspace allows employees to find and use internal information more easily by creating
custom Al agents that can access and understand data from various enterprise sources.
D. Agentspace is mainly designed for building and deploying custom machine learning models for
predictive analytics.
C
A. Data analysis agent
B. Workflow agent
C. Data agent
D. Code agent
D
A. Fine-tune the underlying language model with a broader dataset of general knowledge.
B. Increase the temperature setting of the language model.
C. Implement grounding techniques.
D. Reduce the token count parameter.
C
A. Unlabeled data
B. Labeled data
C. Structured data
D. Raw data
B
A. To generate realistic images of phone cases on devices from text descriptions of designs.
B. To analyze customer feedback to identify popular phone case design trends.
C. To predict demand for different phone case designs based on sales data.
D. To transcribe customer audio feedback on prototypes of phone case designs.
A
A. Gemini
B. Gemma
C. Veo
D. Imagen
D
A. Veo generates videos from static inputs like text or images and cannot process or dynamically
visualize live data.
B. Veo may lack specific scientific visualization styles needed for accurate data representation.
C. Veo use for this scenario would require too many computational resources.
D. Veo is designed for short-form video, not continuous, long-duration live data displays.
A
A. To generate realistic images of phone cases on devices from text descriptions of designs.
B. To analyze customer feedback to identify popular phone case design trends.
C. To predict demand for different phone case designs based on sales data.
D. To transcribe customer audio feedback on prototypes of phone case designs.
A