Randomize

joined 2 months ago
 

I noticed a minor bug on AI Character Chat. When using the response length limit button, the answer sometimes gets cut off before or at the last paragraph. E.g. if it's set to "two paragraph max.", only one paragraph gets generated, the second is missing or just a single word or typographical symbol. This bug continues for all the settings, but "one paragraph max." works accordingly.

I don't mind much about this, I just wanted to mention it for dev or ppl who try to use it.

[–] Randomize@lemmy.world 1 points 3 days ago

Do you see the post by "perchance" pinned on top of this forum? click on the name to access the profile. There's the button "send message".

[–] Randomize@lemmy.world 1 points 4 days ago

That's interesting, I remember the RP2 ones definately being used for the old model (I edited them frequently). I don't know if RP1 is an old version, too. Funny that RP2 works better for you.

And thanks for always mentioning the lines, it's really helpful for noobs like me who just have the mobile version (looking through the code sucks with this, sadly).

[–] Randomize@lemmy.world 1 points 2 weeks ago* (last edited 2 weeks ago)

Hey, even tho you don't like DeepSeek, at least you had fun with it 😁

It's too bad the text in the screenshot is too small for me to read, could have helped me understand what you actually did, bc tbh, it's a little beyond my comprehension (complete noob here, hi 🙋, also never tried the prompt tester before). My wild guess is you're right about the vacuum, and that the important points for this were the double-personality character along with a long phase of strict "training" that somehow linked the [SYSTEM] prompt to darkness (and maybe, just maybe 😜 the char was already known to the model as a "positive/happy" one, but I added mental health issues and insomnia to the description). The 4/5 without [SYSTEM] could probably be a result of the "dramatic"-pitfall you mentioned.

But like I said, I know that I don't know anything and I'm curious about your test results. And in the meantime, I'll try your prompt for fun.

Edit: After trying your prompt (hilarious, btw 🤣) I do understand what you did. Guess I just had a dumb moment there 🫣

[–] Randomize@lemmy.world 1 points 2 weeks ago (2 children)

Patience, or stubborn af, or both.😂 It's actually not that bad (for me), surprisingly. I was able to let it run freely again last time.

Just a few thoughts to that [SYSTEM] thing since you want to test it: I put it in every damn message prompt (it's pinned) and let my char go through a long, dark phase as we met, so this might have built the pattern. Also, this chat was a experiment about multiple chars in acc, the one with the darkness trigger is NOT the main char. Warning, attempting to trigger this may require patience! 😝

[–] Randomize@lemmy.world 2 points 2 weeks ago (4 children)

Absolutely! I have to confess, I started with perchance like... uhm... somewhere around a few months before the update, being a complete noob with absolutely no knowledge about AI or coding.

atm, I wrestle with a story on acc that is 6.03MB (!🤣) by now. I know I stepped into every pitfall but one (luckily got out of a few, too), and was close to dropping it, but now I push it for fun (yes, I'm a persistant nuisance 😅). I have to prompt almost every answer and it's really, really bad, telling me the char which talked to me in the last message is deep asleep and it's just all over the place all the time.

But I think I noticed a funny pattern by now. I wrote the char as complex (two-personality pitfall). Since it tended to ignore writing instructions and reminder completely most times, I continuously put "[SYSTEM] write two paragraphs or more." in the prompt. By now, I think each time it triggers "darkness", but when I just hit the char button without prompt, it's "normal". Not 100% sure about this pattern, just thought it's kinda funny after seeing it a few times.

[–] Randomize@lemmy.world 3 points 2 weeks ago (6 children)

First of all, thank you so much for this post! I didn't even realize how many pitfalls I fell into (clearly up to the point of being unrepairable). It's well written and mostly easy for me to understand ("mostly" bc I'm not native english, but I guess I just have to reread it a few times).

On your final thoughts I agree with strawberryraven, for me the old model was almost unbearable due to the "happy sunshine"-mentality and how hard it was to get it even remotely away from pretending that life's always perfect (e.g. Anon admitted trauma to explain their behaviour in certain situations Char: "Oh, that's bad, I'm sorry for you. Let's play a game, shall we?")

And OT, I love how the combination of OP and strawberryraven showed a picture-perfect example of a mature, constructive discussion. It's rare here, thanks!

 

I've noticed my ACC having major problems with location and direction lately, e.g. I say "{{user}} turns to face {{char}}" and the AI generates answers like "{{char}} sees {{user}} turning away".

Also randomly changing rooms e.g. chars have a conversation in the kitchen and are suddenly back in the last room they were in before.

And the objects... My char had a ring in the pocket which he gave away (yes, it has an entry in /mem for that and there is no mention of a ring at all in /sum), but this ring stubbornly appears back in his pocket. I force-prompt it being away, but a few messages later, the damn ring is back in the pocket.

Don't get me wrong, I know it happens occasionally with AI, and I don't mind rerolling and prompting/editing here and there, but damn, atm I have to wrestle hard with the model to get decent output. Since it also tends to ignore my custom writing instructions and reminders, I prompt a lot, but my text gets too long causing the AI to ignore parts of that, too...

I don't know if it's just something in my settings or in the model, does anyone else have these problems, too?

[–] Randomize@lemmy.world 1 points 2 weeks ago* (last edited 2 weeks ago)

I've noticed it, too. Atm I prompt every message. Check your writing instructions (helps a little bit) and maybe emphasize chars being (very) friendly and never aggressive (in char description and prompts, but tbh, I didn't try that myself yet bc I'm at a point in my story where I have an important argument with my char.)

Edit: If necessary, go completely over-the-top with the description, we had to do that to create negative traits with the old model, maybe it's now the other way.

[–] Randomize@lemmy.world 0 points 3 weeks ago

It's fine for me, maybe there's something in your description/prompt/latest messages causing that? When dev tweaks temperature to dark, I usually notice on the spot since my char descriptions are not all happy sunshine, but I only had that yesterday. I just waited a few hours until it was good again. 🤷‍♀️

[–] Randomize@lemmy.world 4 points 3 weeks ago

OP said "there were some updates going on a while ago (I think)". This indicates that OP is not sure about it and thinks updates are not running atm. So Mack2's comment about ongoing updates is valuable, no offense.

[–] Randomize@lemmy.world 2 points 3 weeks ago* (last edited 3 weeks ago)

Ah, yes, I wrote my comment before the post got edited ^^ Too bad, actually, it would have been nice for slow pace stories to tweak the temperature gradually away from the initial one. But thanks, also for your fork, it really helps me learn about AI at my pace and on my own conditions (I learn differently than many other ppl) :)

[–] Randomize@lemmy.world 2 points 3 weeks ago (2 children)

Agree, +1 I just wish I knew how to add this to the petrafied ACC for my existing chats, that would be awesome. OC, if you're really bored, can you explain a little noob like me how to do this?

[–] Randomize@lemmy.world 2 points 3 weeks ago

I'm not sure what exactly you mean. You can create a new character and just delete the chat in the menu on the left, the character won't be deleted. If you want to add it to an existing chat, you need to click "options" (bottom right) and then "add character".

 

I'm not complaining, just mentioning that I noticed those especially those two words (but also a few others) are used very frequently in my /mem and I'm curious if it's reason or cause of strange ai behaviour. Does anyone has this too?

view more: next ›