Banned words.


The AI has a function to ban words globally or in an adventure. I responsibly declare that this function absolutely does not work. Banned words still do not disappear from the text, and sometimes even on the contrary, the Ai abuses their use.


Activity Newest / Oldest



Status changed to: Closed


Twelfth Wayfare

I find that the global word ban works *most* of the time, and it usually fails when the AI gets really hung up on a word or term (around last December I couldn't get the game to stop bringing up "elfstones," and lately the game's really hung up on "THE END"). The adventure ban simply doesn't work at all - it prevents the *user* from entering certain words, but it doesn't stop the AI whatsoever.

  • Avatar


oh so the the end is happening due to it being banned-



Nah, a global ban is totally not working as well... I just have an idea how this function should work, because I used a similar one in NovelAI.
Novel has it almost completely blocking specified words. The funny thing is, when their AI really wants to insert banned word, it will make a mistake on purpose and distort the word (Luke--> Lukke, Lyke, Ly~ke). :)

Earlier, this AID function worked as a multiplayer option, which prohibited users to use banned words... From that perspective - it works as it should...
It's just that, after they implemented filtering, I saw somewhere that this feature should also block what the AI generates... And their FAQ hints at it.
Maybe it's just another misunderstanding?



Yeah. Banned words completely don't work.
I noticed this quite long ago, but forgot to write about it.
An example I had encountered just yesterday:
I had a spot in a scenario where the AI was introducing a character's name, and when I pressed 'Retry' - it would generate Kid (Sir Kid) pretty often. After I added 'Kid' as the banned word - nothing changed at all.
P.s. I use Griffin, the traveler sub. right now.



Status changed to: Planned