Cognitive Prompts

 Structure a prompt this way to aid  the AI. 

Goal declaration, Decomposition, Filter, Reorganization, Pattern recognition, Abstraction,  Generalization, and integration. 

1. Goal

   Declare the overall goal, possibly with all components declared within the sentence or paragraph. Goal = A

2. Decomposition

   decompose the goal into steps, and declare the steps as an ordered list. A = {p1, p2, ..., pn}

3. Filter

   Keywords section. State keywords selected from the previous two sections. Filter(p) is element of pn.

4. Reorganization

   The aim is to view the steps and goal through different perspectives and state relationships differently, as you would manipulate algebraic equations to create new but equivalent equations. p1 = {d1, ... , dL}

5. Pattern Recognition

   declare any patterns in any Data including in your prompt, this helps the machine to shortcut, dL = {mp1, mp2, ... mpn}

6. Abstraction

   Declare the broad abstractions concerning what is being made. Abstractions are not yet generalization . Abstractions are like shared hyponyms. This helps the machine to create its ontology concerning your prompt. dL = {hn1, ... hnm}

7. Generalizations

   Given the Reorganizations, you can declare generalized rules and principles that will apply. This can be like Variable range declarations as well. declare the application of abstractions on your steps. A = { {hn1 + hnm},{hn2 ++ hnm} }

8. Integration

   Create a new solution Q to the goal A. Q = {DL, ..., D0} = A = { {Hn1 + hnm}, ...} = {p1, p2, ..., pn}

https://arxiv.org/html/2410.02953v2