Forum Replies Created
-
AuthorPosts
-
March 13, 2023 at 8:25 am in reply to: duplicate headers #7031
Have created a handful of posts and so far so good, thank you!
-
March 13, 2023 at 12:59 am in reply to: duplicate headers #7028
Sounds good. So this is still in development? I’m still seeing duplicate titles quite a bit which makes it tough to trust any bulk processes
-
March 11, 2023 at 8:50 am in reply to: duplicate headers #7026
Tried to experiment more after the recent update. I think this issue might have to do with the Writing Tone? When I used the default “informal” duplicate titles seemed very very rare. But when I switched to “humorous” it was happening most of the time.
Are you able to test to confirm? I have no idea why the Tone would matter, but when I switched back and forth between various settings, this seemed like one of the variables that was affecting results.
Thanks.
-
March 10, 2023 at 4:58 pm in reply to: duplicate headers #7024
Ok will try it. Have you been able to reproduce the duplication issue?
-
March 10, 2023 at 12:02 pm in reply to: duplicate headers #7022
I also don’t think it should be numbering these headers like that. It should just give them an h2, but they don’t need to be numbered (#1, #2 etc).
-
March 10, 2023 at 11:59 am in reply to: duplicate headers #7019
-
March 10, 2023 at 10:56 am in reply to: duplicate headers #7017
I mean duplicate headers*
-
March 10, 2023 at 10:56 am in reply to: duplicate headers #7016
Still getting duplicate titles unfortunately 🙁
-
March 10, 2023 at 12:59 am in reply to: article length issue & formatting #7008
Davinci models seem to suffer the same issue unfortunately
-
March 9, 2023 at 8:57 am in reply to: article length issue & formatting #7001
Sounds good. Are you also working on optimizing the PAA feature at all? As is, it works pretty well, but many times the responses/answers begin with the AI saying that “it’s not human”, or “it can’t provide such a personal response”, etc.
Few examples provided in the attached screenshots. I’m thinking if the underlying code/prompt was adjusted a bit, we might be able to prevent responses that make it very obvious the content is being AI-generated. Poor responses like this make it very difficult to trust the quality of the content that’s being created.
I’m also hoping that maybe you can implement the ability for users to customize or add sections to the bulk API form? If we could add multiple custom sections for prompting then it would likely be easier to generate longer articles.
Looking forward to seeing the next update, thanks.
Attachments:
You must be logged in to view attached files. -
AuthorPosts