Prompt Engineering QUIZ (MCQ QUESTIONS AND ANSWERS)

Deleting old prompts

Using only the latest prompts

Hardware versioning

Tracking and managing different iterations of prompts for optimization

Combining reasoning and acting by having the model alternate between thought and action

Making the model react emotionally

Quick responses only

Passive responses

Using the same prompt repeatedly

Model talking to itself

Generating multiple reasoning paths and selecting the most consistent answer

Checking grammar only

A prompt about prompting

A prompt that instructs the model how to create or improve prompts

A deleted prompt

A hardware setting

To limit creativity

To confuse the model

To get structured responses (e.g., JSON, bullet points, tables)

No specific purpose

Being rude to the model

Using negative numbers

Explicitly stating what you don't want in the output

Deleting prompts

Deleting prompts

Breaking complex tasks into smaller, manageable sub-prompts

Combining all prompts into one

Encrypting prompts

Making tasks harder

Confusing the model

No advantage

Guiding the model to produce outputs within desired boundaries (length, tone, format)

Using a single prompt

Using multiple variations of prompts and combining their outputs

Deleting all prompts

Translating prompts

Clarity doesn't matter

Vague prompts work better

Clear prompts lead to more accurate and relevant responses

Only for formal writing

Learning during model training

Learning from external databases

No learning occurs

The model learning from examples provided within the prompt itself

Being vague

Using general terms only

Keeping prompts very short

Providing detailed, precise instructions to reduce ambiguity

Drawing tree diagrams

Exploring multiple reasoning paths in a tree-like structure before reaching a conclusion

Planting trees

Random branching

Assigning a specific identity or expertise to the model to shape responses

Creating fictional characters

Personal data collection

User authentication

Hardware calibration

Adjusting prompts based on model responses to improve accuracy

Temperature settings only

Deleting prompts

Helping the model understand complex concepts through familiar comparisons

Making prompts longer

Confusing the model

No benefit

Providing both positive and negative examples to clarify expectations

Using dark mode

Comparing models

Deleting contrasts

Deleting information

Conveying the same meaning with fewer tokens to save context window space

Making prompts harder to read

Increasing prompt length

A racing competition

Random access control

A programming language

Role, Action, Context, Expectation - a structured approach to prompt creation

Using the same prompt once

Deleting prompts repeatedly

Refining prompts through multiple attempts based on previous outputs

Random prompt generation

Following a single best path

Random searching

Tracking multiple possible sequences simultaneously

Backward searching only

The model is trained on zero data

The model can perform tasks without specific fine-tuning or examples

The model outputs zero results

The model only learns from images

No examples provided

Providing ten examples

Providing one example to guide the model

Providing hundreds of examples

The model can be guided with just a few examples

No examples are needed

It is always slower

It only works for mathematics

Asking the model to think randomly

Asking the model to show step-by-step reasoning

Shutting down the model

Forcing the model to answer quickly

An error message

A debugging tool

A hardware setting

Initial instructions that define the model's behavior

The physical temperature of the model

The randomness/creativity of the output

The training speed

The model size

Selecting from the k most probable tokens

Selecting only one token

Random selection

Always selecting the least probable token

Selecting only the top token

Selecting tokens up to a cumulative probability threshold

Random selection

Always selecting the last token

Random token selection

Selecting less probable tokens

Always selecting the most probable token

Not selecting any tokens