Chain Decorator¶

A meta-decorator that applies multiple decorators in sequence, with each decorator processing the output of the previous one. This enables complex transformations by combining multiple simpler decorators in a pipeline.

Category: Meta

Parameters¶

Examples¶

Basic sequential application of decorators¶

+++Chain(decorators=[StepByStep,Concise])
Explain how neural networks learn.

First generates a step-by-step explanation of neural network learning, then applies conciseness to that output

Complex decorator chain with visible intermediate steps¶

+++Chain(decorators=[Socratic,Academic,TreeOfThought], showSteps=true, stopOnFailure=false)
Discuss the ethics of autonomous weapons.

Shows the progression of applying Socratic questioning, then academic tone, then tree-of-thought reasoning to discuss autonomous weapons ethics

Model-Specific Implementations¶

gpt-4o¶

Instruction: Apply these decorators in sequence: {decorators}. Each decorator processes the output of the previous one. {showSteps} {stopOnFailure}

Notes: This model handles decorator chains well, but benefits from clear instructions about showing steps or handling failures

gpt-4-turbo¶

Instruction: Process this request through a series of steps. For each step, apply one of these decorators in this exact order: {decorators}. Make sure to follow the exact order and treat each output as input to the next decorator. {showSteps} {stopOnFailure}

Notes: This model may need more explicit instructions to maintain the correct sequence of decorators

Implementation Guidance¶

Sequential application of step-by-step and concise decorators¶

Original Prompt:

Explain how neural networks learn.

Transformed Prompt:

Please process this request through a series of transformations, applying each decorator in sequence. Apply the following decorators in this exact order: StepByStep, Concise. Each decorator should process the output of the previous one. Only show the final output after all decorators have been applied. If any decorator in the chain cannot be applied correctly, stop the processing and explain the issue.

Explain how neural networks learn.

Complex decorator chain with visible intermediate steps¶

Original Prompt:

Discuss the ethics of autonomous weapons.

Transformed Prompt:

Please process this request through a series of transformations, applying each decorator in sequence. Apply the following decorators in this exact order: Socratic, Academic, TreeOfThought. Each decorator should process the output of the previous one. Show the intermediate outputs after each decorator in the chain, clearly labeling each step. Continue processing even if some decorators cannot be fully applied, making best efforts for each step.

Discuss the ethics of autonomous weapons.

Transformation Details¶

Base Instruction: Please process this request through a series of transformations, applying each decorator in sequence.

Placement: prepend

Composition Behavior: override

Parameter Effects:

• decorators:

• Format: Apply the following decorators in this exact order: {value}. Each decorator should process the output of the previous one.

• showSteps:

• When set to true: Show the intermediate outputs after each decorator in the chain, clearly labeling each step.

• When set to false: Only show the final output after all decorators have been applied.

• stopOnFailure:

• When set to true: If any decorator in the chain cannot be applied correctly, stop the processing and explain the issue.
• When set to false: Continue processing even if some decorators cannot be fully applied, making best efforts for each step.

Compatibility¶

• Requires: None
• Conflicts: None
• Compatible Models: gpt-4o, gpt-4-turbo
• Standard Version: 1.0.0 - 2.0.0

Related Decorators¶

• All: Enhances Chain Chain can work with any decorators, as it's designed to combine them in sequence
• Priority: Conflicts with Chain Priority and Chain represent different approaches to handling multiple decorators (parallel vs. sequential)