karsuren

3 ms·

1. 1.▲
   by karsuren 3y ago

   You should probably add an example for using the special commands '.critic' '.resolve' commands that you are projecting as the key 'spice' of your project. There is no 'how to use' or 'how it wor
2. 2.▲
   by karsuren 3y ago

   What is the perf of GPT4 vs GPT3.5 in the 'reasoning', 'reflection', 'critisism' and 'resolver' tasks like in your project? I see that you have commented out gpt3.5 and replaced with GPT4 in config ya
3. 3.▲
   by karsuren 3y ago

   Claude has 100k context for around 2$ per million tokens. With GPT4's 4-8k token limit, anything but very small projects in their early phase can benefit from this. Also GPT4 would be far too cost prohibitive