79478148

Date: 2025-03-01 19:36:04
Score: 2
Natty:
Report link

I'm going to recommend starting with a larger pre-trained model. Since you’ve already experimented with a model in the 6.7B range; you can stick with something that has a larger number of parameters. It should provide more robust performance for handling complex prompts; even though it's not fully tuned to your exact needs.

Reasons:
  • No code block (0.5):
  • Single line (0.5):
  • Low reputation (1):
Posted by: LiquidDevOps