Settings
Preferences and model settings.
Common Settings
System Assistant
Language Model
Text-to-Speech
Default Assistant
About
Powered by
LobeHub
Role Setting
OK
Cancel
Assistant Information
Auto Generate
Avatar
Background Color
Name
Assistant Description
Tag
Enter tag
Chat Settings
Chat Window Style
Conversation Mode
Document Mode
User Input Preprocessing
The user's latest message will be filled into this template
Auto Create Topic
Whether to automatically create a topic during the conversation, only effective in temporary topics
Message Threshold
Automatically create a topic when the current message count exceeds this value
Limit History Message Count
Enable History Message Length Compression Threshold
Model Settings
Model
model
ChatGPT model
OpenAI
GPT-3.5 Turbo
16K
Randomness
temperature
The higher the value, the more random the response
Top P Sampling
top_p
Similar to randomness, but do not change together with randomness
Topic Freshness
presence_penalty
The higher the value, the more likely it is to expand to new topics
Frequency Penalty
frequency_penalty
The higher the value, the more likely it is to reduce repeated words
Enable Max Tokens Limit
Speech Service
Text-to-Speech Service
If using OpenAI text-to-speech service, make sure the OpenAI model service is enabled
Text-to-Speech Voice
Select a voice for the current assistant, different TTS services support different voices
alloy
Speech Recognition Language
The language of the speech input, this option can improve the accuracy of speech recognition
Plugin List
Custom Plugin
