ReAct pattern in Prompt Engineering / GenAI - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to initialize the ReAct agent with the correct reasoning step.

Prompt Engineering / GenAI

agent = ReActAgent(reasoning_steps=[1])

Drag options to blanks, or click blank then click option'

A1

B3

C0

D5

Attempts:

3 left

💡 Hint

Common Mistakes

Setting reasoning_steps to 0 disables reasoning.

Using too many steps initially can complicate the agent's behavior.

The ReAct pattern typically starts with one reasoning step to allow the agent to think and act iteratively.

2fill in blank

medium

Complete the code to add an observation after the agent takes an action.

Prompt Engineering / GenAI

agent.observe(action_output=[1])

Drag options to blanks, or click blank then click option'

Aobservation

Baction_result

Caction_output

Dresponse

Attempts:

3 left

💡 Hint

Common Mistakes

Using a variable name that is not defined or unrelated.

Confusing 'observation' with the variable holding the action output.

The agent observes the result of its action, which is stored in 'action_result'.

3fill in blank

hard

Fix the error in the code to correctly perform the agent's think-act cycle.

Prompt Engineering / GenAI

thought = agent.think(input_text)
action = agent.act([1])

Drag options to blanks, or click blank then click option'

Aagent

Binput_text

Caction

Dthought

Attempts:

3 left

💡 Hint

Common Mistakes

Passing the original input instead of the thought.

Passing the action variable before it is defined.

The agent acts based on its thought, so the argument to act() should be the 'thought' variable.

4fill in blank

hard

Fill both blanks to create a loop that runs the ReAct agent for 3 cycles.

Prompt Engineering / GenAI

for _ in range([1]):
    thought = agent.think(input_text)
    action = agent.act(thought)
    agent.observe([2])

Drag options to blanks, or click blank then click option'

A3

Baction

Cthought

Dinput_text

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'thought' instead of 'action' for observation.

Setting the loop count incorrectly.

The loop runs 3 times, and the agent observes the result of its action each cycle.

5fill in blank

hard

Fill all three blanks to define a function that runs the ReAct agent for a given number of steps and returns the final thought.

Prompt Engineering / GenAI

def run_react_agent(agent, input_text, steps):
    for _ in range([1]):
        thought = agent.think(input_text)
        action = agent.act([2])
        agent.observe([3])
    return thought

Drag options to blanks, or click blank then click option'

Asteps

Bthought

Caction

Dinput_text

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'input_text' instead of 'thought' for the act method.

Observing 'thought' instead of 'action'.

The loop runs for 'steps' times, the agent acts on 'thought', observes 'action', and finally returns the last thought.