Description
This week we had an issue where we exceeded the context window, it would be nice if on LLM requests there was some sort of summary tab where we can see how many input tokens for each request are taken up by
- user message
- assistant messages
- Tool calls
Description
This week we had an issue where we exceeded the context window, it would be nice if on LLM requests there was some sort of summary tab where we can see how many input tokens for each request are taken up by