was reading through open llama, looks like way to get pertinent results is via different ranking algorithm and score based on convergence. then shove that back

was reading through open llama, looks like way to get pertinent results is via different ranking algorithm and score based on convergence. then shove that back into the LLM