NEW
Font size
WorksheetsCCPS Day 2 - Latihan
Total questions: 25
Worksheet time: 13mins
When implementing Chain of Thought (CoT) prompting for a complex business decision, which strategy would be MOST effective in enhancing the reasoning process?
Requesting only the final answer without intermediate calculations
Asking the LLM to enumerate each reasoning step sequentially
Providing multiple unrelated examples to confuse the model
Encouraging explicit reasoning while prompting elaboration on each conclusion
A data scientist wants to analyze customer feedback using the ReAct framework. What is the PRIMARY advantage of this approach over standard prompting?
It eliminates the need for external data sources completely
It alternates between reasoning and action to maintain context relevance
It only focuses on the final output without intermediate steps
It removes the requirement for logical analysis in responses
Which advanced technique would be MOST appropriate when an LLM refuses to answer due to unclear phrasing?
Template Pattern for structured output formatting
Game Play Pattern for interactive content generation
Refusal Breaker Pattern with alternative question wordings
Infinite Generation Pattern for continuous output creation
In the Cognitive Verifier pattern, why does subdividing complex questions improve LLM performance?
Research shows LLMs perform better when answering multiple focused sub-questions
It reduces the total number of tokens required for processing
It eliminates the need for user input during interaction
It automatically generates creative content without guidance
A marketing team needs to create multiple campaign variations. Which pattern combination would be MOST efficient?
Context Manager with Fact Check List only
Infinite Generation with Template Pattern for automated variations
Question Rewriting with Refusal Breaker exclusively
Meta Language Creation with Cognitive Verifier alone
What is the PRIMARY purpose of the Context Manager pattern in advanced prompting?
To generate unlimited outputs without user intervention
To create alternative languages for prompt communication
To specify or remove contextual boundaries for focused conversations
To automatically verify factual accuracy of all responses
When using the Recursive Reprompting and Revision (Re3) technique, what distinguishes it from standard iterative prompting?
It only works with simple, single-step tasks
It requires manual revision after each generation
The model self-evaluates and revises content during generation
It cannot be used for comprehensive document creation
A developer wants to implement the Flipped Interaction pattern for gathering requirements. What is the KEY characteristic of this approach?
The LLM asks questions to obtain information instead of answering
The user provides all information upfront without interaction
The conversation follows a strict predetermined script
The LLM generates content without any user input
In which scenario would the Meta Language Creation pattern provide the MOST value?
When conventional language adequately expresses all required concepts
When creating a specialized command system for repetitive tasks
What is the MOST critical consideration when implementing the Fact Check List pattern?
Placing the fact list before the main output for visibility
Eliminating all need for user verification of information
Allowing users to understand statements before seeing verification needs
Reducing the number of facts to minimize user workload
When combining Question Rewriting with other patterns, which combination would BEST address unclear user queries about productivity?
Question Rewriting with Game Play Pattern exclusively
Question Rewriting with Cognitive Verifier and domain-specific context
Question Rewriting with Infinite Generation only
Question Rewriting without any additional pattern support
A content creator needs to generate long-form articles with consistent quality. Which technique ensures continuous self-improvement during generation?
Template Pattern with fixed structure requirements
Context Manager with strict topic boundaries
Recursive Reprompting and Revision (Re3) with self-evaluation
Infinite Generation without quality control mechanisms
In the Recipe pattern, what is the PRIMARY function when users know their end goal but not the exact steps?
To provide only the ingredients without any sequence
To create new goals based on available resources
To output a complete sequence while identifying missing or unnecessary steps
To eliminate all user input requirements from the process
When implementing the Template pattern, what is the MOST important instruction to give the LLM?
To freely modify the template structure for creativity
To ignore placeholder positions in the template
To follow the precise template structure without rewriting it
To generate templates automatically without user specifications
A game developer wants to create educational content with varying scenarios. Which pattern combination would be MOST effective?
Fact Check List with Context Manager exclusively
Game Play Pattern combined with Visualization Generator*
Refusal Breaker with Meta Language Creation only
Recipe Pattern with Template Pattern alone
What distinguishes Meta Prompting from other advanced techniques?
It focuses solely on answering user questions directly
It involves asking the LLM to create, refine, or evaluate prompts
It eliminates the need for any prompt engineering skills
It only works with basic conversational interactions
In intermediate prompting, when should the Extracting technique be applied?
When generating creative content without source material
When expanding brief outlines into full documents
When identifying and organizing specific information from complex texts
When transforming content tone without changing meaning
A business analyst needs to infer sentiment from customer reviews. Which intermediate prompting technique is MOST appropriate?
Extracting specific data points from structured reports
Inferring emotions and sentiments from unstructured text
Transforming language between different formats
Expanding bullet points into comprehensive paragraphs
When using the Transforming technique, what type of modifications can be performed?
Only language translation between different spoken languages
Exclusively correcting spelling errors in documents
Multiple changes including tone adjustment, format conversion, and grammar correction
Only generating new content without modifying existing text
What is the KEY benefit of using the Iterative prompting approach in intermediate techniques?
It eliminates the need for any user feedback
It generates perfect outputs on the first attempt
It allows continuous refinement based on feedback and experimentation
It prevents any modifications to initial prompt structures
In the Expanding technique, why is it recommended to start with an outline before full content?
It reduces the total time required for generation
It helps overcome token limits through chunking while maintaining structure
It eliminates the need for any contextual information
It automatically generates all content in a single pass
When implementing the Summarizing technique, what should be the PRIMARY focus?
Expanding the content to include more detailed information
Transforming the tone without changing the content
Condensing information while retaining essential key points
Extracting only numerical data from the source text
A researcher wants to ensure factual accuracy in LLM-generated content. Which pattern should be prioritized?
Infinite Generation for multiple content variations
Game Play Pattern for interactive engagement
Fact Pattern for accuracy
When would the Context Manager's "Start over" command be MOST valuable?
When continuing with the same topic requires more depth
When expanding on previously discussed related concepts
When completely shifting to an unrelated topic requiring fresh perspective
When maintaining consistency with all previous conversation context
In applying Chain of Thought (CoT) prompting to a troubleshooting scenario, what should the prompt structure emphasize?
Only the final solution without diagnostic reasoning
Multiple unrelated problem-solving approaches simultaneously
Step-by-step diagnostic process with explicit reasoning at each stage
Immediate conclusions without intermediate analysis steps
