Chat Inference
Last updated
Last updated
The chat interface of GenAI Studio is where you will spend most of your time. Familiarizing yourself with its basics is crucial as it features intuitive and familiar design elements. Updates to the interface may include additional icons for improved functionality.
User messages are those you send. They are used to find similar documents and are sent to the LLM for processing.
Copy: Copy the content of the message.
Edit: Amend and automatically resubmit the conversation from that point to the LLM. Note that this truncates all messages below the edited content.
Speak: Use native text-to-speech modules, OpenAI Voice, or 11Labs voice to vocalize your text.
LLM messages are responses from the active LLM in the chat session. These responses help find similar documents and contribute to future conversations. History management is automatic when the context window is exceeded.
Copy: Copy the content of the message.
Edit: Modify the output of an LLM message for correctness without resubmitting the prompt. This simply updates the history.
Regenerate: Resend the same prompt and history to the LLM for a new answer.
Feedback (Thumbs Up & Down): Provide qualitative feedback on LLM responses. Feedback doesn’t affect message history or future responses but is useful for exporting chats to create fine-tunes.
Slash Commands: Inject standard text into your prompt using short-keys. Create and manage commands here.
Default Slash Commands: Built-in commands with special functions like /reset.
@Agent Invocation: View available @agents and their skill sets. Starting a prompt with @agent initiates an agent session.
Font Size: Set the default font size for your profile.
Microphone: Enable voice-to-text inputs for your LLM prompts (not available on Desktop).
