Does anyone know how good settings for this Proxy?
Hello! I'm currently using hermes-3-llama-3.1-70b and right now whenever I type, it seems to talk for my character. However, it only seems to be when I use hermes-3-llama-3.1-70b. Does anyone know the correct generation settings to stop this? If not that is okay, I'll trial and error through lol but I was just wondering if anyone knew. I'm also using it through NanoGPT since Openrouter doesn't allow payment in the UK.
My Generation Settings:
Temperature: 0.75
Max Tokens: 200
Context Size: 50000
Top K: 0
Top P: 0.9
Rep Penalty: 1.1
Freq Penalty: 0.4