search

jianyangg / local-llm

DSTA Internship Project
0 stars 0 forks source link

Feedback loop #20

Open jianyangg opened 3 days ago

jianyangg commented 3 days ago

For the agentic workflow, we are only relying on the temperature setting of the llm to generate a different response. Instead, we want the agents to give and receive feedback so that answer can be evaluated by each grading agent which may improve the answer iteratively.

hanchingyong commented 2 days ago

still want do this? can be a good to have bah