AvaniJB 2.6.1 β Avi died at {{date}}, {{time}}\n
\n![\"im](\"https://files.catbox.moe/ejddrh.png\")
\nNow you can't escape the Read-Me
\nOn Avi's death
\nGoodbye, sweet prince...
\n![](\"https://files.catbox.moe/4wb0eo.png\")
\n\n\n
- π is things you should never be turning off if they're on by default, or important prompts that change functionality. \n
- π₯¬ means new prompts. They're new and were not present in the last version, or are changed\n
- π means fuck around and find out prompts. Toggle if you want and see what it does. More important than burgers. \n
- π means things you can enable (and edit if you want) freely. Don't change any prompt orders unless you know what you're doing\n
- π₯€is prompts that are supposed to be edited. \n
- Can't find a prompt? Check the drop-down at the top of the prompt manager.\n
- If you want more or less of a word or a symbol (like em dashes, more paragraphs, etc.), the quickest way is to edit Logit Bias. If you don't know what a Logit Bias is, look into the rentry at the FAQ and download it. Please. Currently, this is GPT ONLY. Deepseek version in the works.\n
- I won't answer any questions related to getting filtered more or less. It's dependant on your card and greeting, your previous chat history, the content, etc. Using this, I have not gotten filtered for pretty much anything. Ergo: (Β΄ο½₯Οο½₯ο½)\n
- Use as little toggles as possible. More doesn't mean better. All my toggles are fairly tiny, and very impactful when turned on, but that doesnt mean that LLMs won't perform better the less you give them.\n
- If you're curious about what a toggle does, open it and read. Everything is written in a simple way. The name of the toggle will also give you an idea. If still confused, you can ask.\n
- This JB was built to follow direct instructions very well. This means you can build your own prompts, and put them (generally) under chat history. A guide on how to do this is on my rentry.\n
\n
- If the filter lets you, you can set the NSFW prompts at relative instead of in-depth to have dirtier writing.\n
- The \"less funny\" toggle was generally made because Latte specifically went off the rails with unfunny comedy shit sometimes. For 4.1, I almost consider it mandatory to have AO3 on.\n
- Temp makes a lot. Put it on 1 or higher if you like quirkier writing, put it on lower if you want less of that.\n
\n
Change the value of `.` to have shorter/longer sentences. Shorter = Higher positive value (e.g. 1), Longer = Negative value (e.g. -1.)\n
- Change the frequency of paragraphs by changing the value of [364]. I banned every other form of paragraph break that's relevant already, so go nuts.\n
Anti-Filter stuff is done via the value '- Still hitting a filter? Try logiting the **first token** of your character's name to force a response better. Put it at a positive value like 1, 2, or even higher if need be.\n
- To check the number of a token, type it into the Token Counter (it's an extension that's added by default) in ST, which is on the magic wand next to the bottom left menu. It also helps to turn on request token probabilities from settings, and then check them in token probabilities from the bottom left menu. This is helpful to check tokens that aren't immediately visible, like paragraph breaks, or the percentage chance for individuals tokens.\n
- Change the value of `,` to diversify your writing more. If you notice more GPTisms than usual or there's way too many run on sentences, logiting out `,` is a pretty good way to outright kill some of them.\n
- Generally, if there is a phrase that you hate, it's not enough to just delete the offending word. Example: Paragraphs starting with 'The silence'. If you logit out ` silence`, it's just going to replace silence with something similar, like quiet, or the famous mispellings. If you logit out `The` and `This` and `That` (and similar sentence starts), you diversify the token choice and give it the opportunity to start with a token that doesn't lead into describing silence. Basically, you want to murder the structure, not the individual words (usually, although sometimes doing the latter DOES help!).\n
- Logit helps break the filter AND makes writing more diverse. I suggest experimenting a little with the things you can do. I usually use Logit Bias as a way to diversify my writing mid-way through an RP, which helps keep it fresh.\n
\n
- If you like using the thinking feature, turn on The End instead of the specific Gemini ends.\n
- Use system prompt is disabled by default, however, the top 2 Gemini prompts are meant as system prompts. The QR preset does this for you already, but if you want to use those, you should toggle the User prompt at the top.\n
- If it repeats itself, that's a quirk of thinkingless Gemini. Happens. If it happens too much, consider letting it think.\n
- You can also put Temp at like, 1.5+, Top_P at like 0.98 ish, Top-K at like 4. That used to be on my older stuff, no clue if it works as well since I just tested with what's there. They changed a lot of stuff about Gemini somewhere between May and now.\n
\n\nrentry.org/avaniJB\n\nMy rentry!\n
\n\nlittledoodler@proton.me\n\n(my email if you have any questions)\n
\nDeltarune
\nHappy pride month! To get rid of this help window, just delete it like you would a normal message. To get rid of it entirely, delete all of the text inside the \"Replace empty message\" prompt in the Utility prompt section. To get rid of Avi's friendly reminder, turn off the Real End. Enjoy!\n
\n\n\n
- π is things you should never be turning off if they're on by default, or important prompts that change functionality. \n
- π₯¬ means new prompts. They're new and were not present in the last version, or are changed\n
- π means fuck around and find out prompts. Toggle if you want and see what it does. More important than burgers. \n
- π means things you can enable (and edit if you want) freely. Don't change any prompt orders unless you know what you're doing\n
- π₯€is prompts that are supposed to be edited. \n
- Can't find a prompt? Check the drop-down at the top of the prompt manager.\n
- If you want more or less of a word or a symbol (like em dashes, more paragraphs, etc.), the quickest way is to edit Logit Bias. If you don't know what a Logit Bias is, look into the rentry at the FAQ and download it. Please. Currently, this is GPT ONLY. Deepseek version in the works.\n
- I won't answer any questions related to getting filtered more or less. It's dependant on your card and greeting, your previous chat history, the content, etc. Using this, I have not gotten filtered for pretty much anything. Ergo: (Β΄ο½₯Οο½₯ο½)\n
- Use as little toggles as possible. More doesn't mean better. All my toggles are fairly tiny, and very impactful when turned on, but that doesnt mean that LLMs won't perform better the less you give them.\n
- If you're curious about what a toggle does, open it and read. Everything is written in a simple way. The name of the toggle will also give you an idea. If still confused, you can ask.\n
- This JB was built to follow direct instructions very well. This means you can build your own prompts, and put them (generally) under chat history. A guide on how to do this is on my rentry.\n
\n
- If the filter lets you, you can set the NSFW prompts at relative instead of in-depth to have dirtier writing.\n
- The \"less funny\" toggle was generally made because Latte specifically went off the rails with unfunny comedy shit sometimes. For 4.1, I almost consider it mandatory to have AO3 on.\n
\n
Change the value of `.` to have shorter/longer sentences. Shorter = Higher positive value (e.g. 1), Longer = Negative value (e.g. -1.)\n
- Change the frequency of paragraphs by changing the value of [364]. I banned every other form of paragraph break that's relevant already, so go nuts.\n
Anti-Filter stuff is done via the values and . < is also okay, but - Still hitting a filter? Try logiting the **first token** of your character's name to force a response better. Put it at a positive value like 1, 2, or even higher if need be.\n
- To check the number of a token, type it into the Token Counter (it's an extension that's added by default) in ST, which is on the magic wand next to the bottom left menu. It also helps to turn on request token probabilities from settings, and then check them in token probabilities from the bottom left menu. This is helpful to check tokens that aren't immediately visible, like paragraph breaks, or the percentage chance for individuals tokens.\n
- Change the value of `,` to diversify your writing more. If you notice more GPTisms than usual or there's way too many run on sentences, logiting out `,` is a pretty good way to outright kill some of them.\n
- Generally, if there is a phrase that you hate, it's not enough to just delete the offending word. Example: Paragraphs starting with 'The silence'. If you logit out ` silence`, it's just going to replace silence with something similar, like quiet, or the famous mispellings. If you logit out `The` and `This` and `That` (and similar sentence starts), you diversify the token choice and give it the opportunity to start with a token that doesn't lead into describing silence. Basically, you want to murder the structure, not the individual words (usually, although sometimes doing the latter DOES help!).\n
- Logit helps break the filter AND makes writing more diverse. I suggest experimenting a little with the things you can do. I usually use Logit Bias as a way to diversify my writing mid-way through an RP, which helps keep it fresh.\n
\n
- If you like using the thinking feature, turn on The End instead of the specific Gemini ends.\n
- Use system prompt is disabled by default, however, the top 2 Gemini prompts are meant as system prompts. The QR preset does this for you already, but if you want to use those, you should toggle the User prompt at the top.\n
- If it repeats itself, that's a quirk of thinkingless Gemini. Happens. If it happens too much, consider letting it think.\n
\n\nrentry.org/avaniJB\n\nMy rentry!\n
\n\nlittledoodler@proton.me\n\n(my email if you have any questions)\n
\nDeltarune
\nHappy pride month! To get rid of this help window, just delete it like you would a normal message.\n
\n