Playing AI Dungeon
Advanced AI Settings
Advanced Settingsare a premium-only AI Dungeon feature that let you customize and control how AI responses are generated during gameplay.
The AI generates the next word by creating a list of all the possible words that could come next and the probability that they would be next, then selecting a word from that list by doing some math with a random number. The Advanced Settings allow you to change the numbers within that calculation that aren't random.
This is complicated, and there are no standards, so it’s highly recommended that you experiment with them and come on the Discord to ask for advice.
You may also find it helpful to read our more advanced article on how the AI works, which will give you a better understanding of how these function beneath the hood.
Raw Model Output
The Interface cuts off everything after the last punctuation mark so that you don’t have a half-finished phrase at the end of the AI’s last generated output. Toggling on
Raw Model Output allows you to see whole, unprocessed outputs from the model, but this may result in incomplete sentences and other issues.
Temperature means how random the AI should be when selecting the next word: lower temperatures make unlikely words less likely to be selected, while higher temperatures will make them more likely to be selected. If the temperature is too high, the AI will start to select some very strange words, but if it is too low, the AI will trend towards the most likely words only, which can become boring and repetitive.
The default temperature is 0.8, though you might consider something like 0.6 if you feel the AI is getting a bit too far from a reasonable response, or you may consider something like 0.95 or even 1.2 if you want to see the AI put out some really uncommon text.
Response Length determines how many words the AI will generate before it finishes. This is a matter of personal preference: You may want short responses that you can quickly read and decide when you want to jump in, or you may want very long responses that you can take your time with. Do whatever fits your play style.
Top-K is only available for Griffin and ChatGPT. It penalizes repetition at the cost of more random outputs.
Top-P affects the randomness of the AI. Higher values mean more randomness. A Top-P of 0.5 to 0.95 is probably best for most users.
If Top-P is too low, AI outputs may be very repetitive and uninteresting. If Top-P is too high, AI outputs may be disjointed and incoherent, possibly even grammatically incorrect.
Repetition Penalty is only available in Griffin. It changes the probability values to make the AI less likely to repeat words. This may sound good, but remember that the names of characters, as well common words like "You", "I", "and", “a”, etc., are going to be penalized, so too high a value will start cutting out these words in favor of some otherwise very uncommon ones, possibly even becoming grammatically incoherent.
Repetition Penalty defaults to zero, and while you can set it very high, the AI will start to trend towards those strange outputs if you put it too far over even 1.
Presence Penalty is only available for Dragon and ChatGPT. It applies a fixed bias only to tokens that appeared at least once.
Count Penalty is only available for Dragon. It applies a bias proportional to the number of times each token that has appeared before.
Frequency Penalty is only available for Dragon and ChatGPT. It is the same as Count Penalty, but divided by the total number of tokens in the text.
Memory Length refers to how much text is sent to the AI every time it’s called. It’s typically best to have this set to the highest value so the AI is always receiving as much context about your Adventure as possible. The context includes information in Memory, any triggered Story Card entries, Author’s Notes, and the most recent player input. Any remaining space in the Memory Length is filled by story text, so maximizing that amount of context the AI receives helps it generate consistently coherent outputs during gameplay.
Tips for Learning Advanced Settings
There are no rules in AI Dungeon
As with anything in AI Dungeon, the best way to get a feel for things is to experiment, and there’s no penalty for doing so. If you are really interested in understanding what these do, the best way to do so is to take the time to change them and see what happens. You can nudge them around based on what you think you'd personally prefer, and you might like it better, or you might go back. Sure, they're going to do weird things with the AI's already weird output, but that's part of the fun of using AI Dungeon!
There are no consequences to messing with these settings
There are no consequence to turning up the Top-P, Temperature, and/or Repetition Penalty to maximum just to see the AI flail about like a tube man. There are no consequence to turning everything down to zero and watching the AI do its best impression of a broken record.
There aren’t agreed upon "best values"
The values these settings default to have tested best in our Train the AI experiments, but they may not work for your style of gameplay, and that’s alright. We’ve provided these settings so players can customize and control their AI Dungeon experience to however is best for them. If you like something that no one else is using, that’s great! At the end of the day, it's your story, your play-through, and your experience, and you are free to do whatever makes you happy with it.
On this page
- Advanced AI Settings
- Raw Model Output
- Response Length
- Repetition Penalty
- Presence Penalty
- Count Penalty
- Frequency Penalty
- Memory Length
- Tips for Learning Advanced Settings
- There are no rules in AI Dungeon
- There are no consequences to messing with these settings
- There aren’t agreed upon "best values"
© Latitude 2023