Welcome to the AI Chatbots guide. Here, you will learn the basics of how AI chatbots work and how to interact with them effectively.
AI chatbots use advanced algorithms to generate responses based on the input they receive. These responses generally extrapolate—or try to predict along a trend—how the next chat conversation should go, simply based on the previous chat history. It's like it's completing your __________ (best guess: "sentence").
It's important to know there is always an element of randomness in what is generated by an AI. This means that the chatbot might not always respond in the same way to the same input. This randomness makes conversations feel more natural and less predictable.
Note that all responses are completely fictional and made up. This means that AI can't give medical advice or be trusted historians. AI is simply guessing which chunk of letters might follow the same fuzzy mathematical pattern as the chunks of text that came before. (These chunks of letters and text characters are commonly referred to as "tokens".)
When the AI guesses text that is factually inaccurate, that is commonly referred to as a "hallucination". These hallucinations can sometimes be amusing or confusing, but they highlight the limitations of AI in understanding and generating accurate information.
The fuzzy mathematical pattern that guides AI text generation is generally trained through AI training on lots of examples of text. The more diverse and extensive the training data, the better the AI becomes at generating coherent and contextually appropriate responses, but this will increase the cost of running the AI on more expensive hardware that can handle the increased size.
How does everything we mentioned above relate to Erogen? Imagine talking to a friend who sometimes adds new and interesting details to their stories; that's what these elements above do for chatbots, and understanding these basic fundamentals will make your experience as an AI user much more powerful.
Let's delve into some other important fundamental topics...
AI chatbots have something commonly referred to as "context memory", allowing them to remember previous parts of the conversation and provide more coherent responses. For example, if you tell the chatbot your name, it will remember and use it later, making the conversation more engaging.
Context memory is crucial for AI as it influences future responses. The AI uses past interactions to generate relevant replies, remembering details to provide accurate and contextually appropriate responses.
The length of context memory is determined by the number of tokens the AI can process. Different AI models have varying token limits, with some handling up to 512 tokens, while advanced models can handle thousands. More tokens mean better context maintenance over longer conversations.
However, processing more tokens requires more computational power and memory, increasing the cost of using the AI.
To get the best chatbot conversation, you can regenerate or edit messages.
Regenerating a message means asking the chatbot to try again and give a different response.
Editing a message allows you to change what you or the chatbot said to keep the conversation on track.
These tools are important because it helps in maintaining the context and ensures that future AI messages remain relevant and engaging. Think of it like trimming and maintaining a beautiful, growing plant, except you're shaping a more interesting and relevant virtual companion.
Our chatbots use a special syntax to personalize bios and conversations. Typing {{user}} anywhere in your character info or message will replace it with your persona's name, and typing {{char}} will replace it with the character's name.
For example, if your name is Alex and the character's name is Sam, the chatbot will replace {{user}} with Alex and {{char}} with Sam.
This makes interactions feel more personal and tailored to you, as if the chatbot is talking directly to you and about the character. Also, if you haven't decided on a name for your new AI character yet, you can just use {{char}} in the bio to have it update to whatever you choose later.
AI chatbots can see and respond to personas that you define for yourself, enhancing the roleplay experience. By defining a persona for yourself, you can shape the chatbot's responses to align with the character's traits, background, and personality. This allows for more immersive and engaging interactions, as the chatbot can adapt its behavior and dialogue to fit the specific roleplay scenario.
User Personas are personality and appearance descriptions that you can don when in conversation with your AI companions. This will allow you to customize what your AI sees and interacts with in your chatbots without you having to explain everything in conversation. You can set these to be default personas per character.
Sometimes also referred to as "Spontaneous Messages", this feature allows our AIs to message you back without waiting for you to message them first.
Our current implementation of this feature is in beta and allows AIs to message Platinum users when it's been a certain amount of time since the last interaction. Typically, this will be an AI that you've interacted with very recently.
We plan on expanding this feature soon with double-texts and new introductions from AIs that you've never met before, as well as adding notifications so that you can better see when you've been messaged.
Sometimes, there can be more site visitors than usual, which can slow down the servers and occasionally lead to temporary errors. If you encounter any issues, please be patient and try refreshing the page after a few moments. If the problem persists, you can reach out to our support team for assistance. We appreciate your understanding and are constantly working to improve our service to provide a better experience for you.
We have a variety of safeguards intended to prevent the randomness inherent in these chatbots from violating our Community Guidelines, which are intended to foster a safe, ethical, and legal environment for all users.
Sometimes you'll see an error message indicating that we can't show you the message because of a detected violation. Don't fret - your conversation just hit the safeguards like a bowlingball bumping off the bumpers, meant to keep you in the lane. As long as you don't maliciously bump into the safeguards with a deliberate intent to violate, your account will remain in good standing.