Show HN: RL Agent that can auto-optimize your LLM prompts
(nomadic-ml.github.io)14 points by varunkrishnan17 4 days ago | 2 comments
Hey everyone! Along with my team, I've developed a reinforcement learning system that automatically optimizes LLM prompts, complete with a visualization feature to track both prompt structure and learning progress over time. Take a look here: https://nomadic-ml.github.io/nomadic/cookbooks/Nomadic_Promp...
Check out our website too:https://www.nomadicml.com/
In terms of how this visualization works: The RL Prompt Optimizer employs a reinforcement learning framework to iteratively improve prompts used for language model evaluations. At each episode, the agent selects an action to modify the current prompt based on the state representation, which encodes features of the prompt. The agent receives rewards based on a multi-metric evaluation of the model's responses, encouraging the development of prompts that elicit high-quality answers.
And see our github repo! https://github.com/nomadic-ml/nomadic
mustafabal 4 days ago | next |
Hi everyone! Fellow co-founder of NomadicML with Varun. We’re fascinated by auto optimization of complex ML systems, and the enterprise benefits it will yield.
Our innovative RL approach to optimize components of your ML systems (not only your LLM but also your RAG, guardrails, prompt tuner, etc…) in production is only but one benefit of the Nomadic Platform. Check out our SDK & Workspace for more!