Wolfram Language Paclet Repository
Community-contributed installable additions to the Wolfram Language
Designing an LLMFunction
opens the ChatGPTPromptExplorerGeneralPalette. This governs the main interactions with this application. |
Opening the ChatGPTPromptExplorerGeneralPalette.
One approach to creating an LLMFunction
First load the package:
Needs["DavidReiss`ChatGPTPromptExplorer`ChatGPTPromptExplorer`"];
From a Chat Notebook
Create a Chat Notebook. For any function keys enclose those in single quotes.
From the click on "Copy ChatObject". Then paste the result in to the LLMFunction[""] expression within a pair of quotation marks
In[33]:=
co=ChatObject
;
|
In[37]:=
llmFun1=LLMFunction[Query[All,"Content"]@co["Messages"]]
Out[37]=
LLMFunction
|
In[36]:=
llmFun1["Pitcher""Whitey Ford","Catcher""Yogi Berra"]
Out[36]=
Pinstripes they donned proud,Yankees legends they became,Ford and Berra's team.
First approach: from a JSON expression
First approach: from a JSON expression
Create a Chat Notebook. For any function keys enclose those in single quotes.
From the click on "Copy Prompts: JSON". Then paste the result in to the LLMFunction[""] expression within a pair of quotation marks
In[31]:=
llmFun2=LLMFunction["[\t{\t\t\"role\":\"user\",\t\t\"content\":\"What team did `Pitcher` and `Catcher` play for?\"\t},\t{\t\t\"role\":\"user\",\t\t\"content\":\"Give your answer as an haiku.\"\t}]"]
Out[31]=
LLMFunction
 | |
|
In[32]:=
llmFun2["Pitcher""Whitey Ford","Catcher""Yogi Berra"]
Out[32]=
Two Yankees greats played,Whitey Ford and Yogi Berra,Bronx Bombers for life.
 |
|
 |
|
""