OVERCOMING LANGUAGE MODEL TOKEN LIMITS WHEN CONVERTING EXECUTABLE CODE

A method of overcoming a token limit of a language model when converting an existing computer executable code to a modified computer executable code. The method includes generating a code segment from the existing computer executable code. The code segment is generated to be within the token limit. The method also includes identifying a prompt that instructs the language model to generate the modified computer executable code. The prompt includes a system command instructing the language model to convert the code segment to the modified computer executable code. The prompt also includes the code segment. The prompt also includes a map defining a mapping between the code segment and the modified computer executable code. The method also includes processing the language model on the prompt to output the modified computer executable code. The method also includes returning the modified computer executable code.