#[βœ… Solved] Parameters

1 messages Β· Page 1 of 1 (latest)

short gazelleBOT
#

For the CosmosRP-3.5-MSR model on ChubAI, here are the recommended settings and parameters to ensure best performance and response quality:

Required Settings

  • API Base URL:
    https://api.pawan.krd/cosmosrp-3.5-msr/v1
  • API Key:
    Enter your Supporter I API key.
  • API Mode:
    OpenAI (Reverse Proxy)

ChubAI Parameter Recommendations

  • Model:
    Set to "gpt-4" or closest available option (ChubAI will map to CosmosRP-3.5-MSR via your base URL).
  • Temperature:
    0.5 (Recommended)
    • Min: 0.3
    • Max: 0.6
  • Max Tokens:
    Use default or up to 2000 if you want longer responses.
  • Presence Penalty / Frequency Penalty:
    Leave at OpenAI defaults unless you want specific behavior.
  • Jailbreak Prompt:
    Enter a single space ( ). Do not use jailbreaks or custom instructionsβ€”they are not needed and can lower quality.

Extra Tips

  • Use OpenAI Default Generation Settings whenever possible.
  • No custom prompts/jailbreaks neededβ€”the model is already uncensored.
  • Make sure your ChubAI is fully updated for best compatibility.
  • If you notice errors or low-quality responses, create a new chat and make sure no leftover jailbreaks or instructions are present.

You can always review specific model info here.

Note: If you encounter any technical issues, you can create a ticket in this channel for direct assistance.

Let me know if you need setup steps for other platforms!

Please note: This is an automated response from @short gazelle. While it strive to provide accurate assistance, It may occasionally make mistakes. If you find any inaccuracies or need further clarification, feel free to wait for our community helpers to give you further guidance. Thank you.

ripe sorrel
#

What about Pre and post History Instructions as well as impersonation prompt

short gazelleBOT
#

[βœ… Solved] Parameters