Raw LLM Responses

If I may ask did you try filtering the scenes/chapters using specific prompts? IE: rewrite in 1st person? ... rewrite in 3rd person close? IE: rewrite in tone XXX? I have been experimenting with Chatgpt (as well as Sage, Claude, GPT4, Dragonly via [poe.com](https://poe.com) ) with already written scenes and transformations of these (including expansion by prompt(s). I takes a lot of testing to get AI to "unwrite" the bland it produces and write in closer to today's standards. You might be able to use some or all of these to write scene\[s\]/chapter\[s\]. Note: This testing was done on GPT3, 3.5 and the "test scene" was 534 words, in 3rd person close (POV of the character). GPT4 is LIGHT YEARS ahead, but testing is much slower at the moment. Note that SAGE, and Claude are also excellent, fast and give fantastic insight into your scene\[s\] and what is missing and how to fix them. Here are some of my test prompts (in order, note ones with "ERROR" did not work or produced minimal results) : 1st: Please review the following passage. If you have recommendations, please create 2 paragraphs per recommendation in the same style as the passage. Keep adjectives to a minimum. Use body language as per show don't tell style. 2nd prompt after “recommendations”: for both recommendations, provide 2 paragraphs based on the same style which can be edited and added into the passage. 3rd prompt: rewrite the paragraphs of each recommendation and use only body language. 4th: rewrite the current versions of recommendations from close 3rd person in Rogue's voice. Rogue is an ai in a human like form grappling with his humanity. 5th: rewrite the last 2 recommendations, make them more vivid and include some of Rogue's thoughts in italics, rather than narrative summary. 6th: ERROR rewrite the last 2 recommendations using body language description - of both Rogue and the other characters. Expand the paragraphs/length of the passages as required to meet this require