IMPORTANT: Scale offers no way to check whether your key is correct. If you see "Valid" here but you still can't send prompts, confirm your key/url.
Context Settings: If you are using GPT-4, you can select up to 7750 tokens of context. Replies will take longer to generate.
This value must match the Spellbook deployment's "Maximum Tokens" setting or you may send prompts that are too large. If you don't know what this means, leave it at 400.
Model Parameters: Scale doesn't support adjusting model parameters from the API.