Arize AI Community Icon

Home
Events

Arize AX Releases
Arize AX Support
Arize News
Arize Observe
Discussions
Introductions
Phoenix Support

Powered by Tightknit

Exploring UI Options for Evaluating LLM Chatbot Performance | Arize AI Community

Exploring UI Options for Evaluating LLM Chatbot Performance

·Nov 26, 2023 05:10 PM

Hey guys I am new to this community. I am building an LLM Chatbot based on some custom knowledge base. I want to use Phoenix Eval to evaluate the Q&A metric. I have tried it it works well. But, is there a UI based version for running evals ? like MLFlow or LangSmith ?

4 comments

· Sorted by Oldest

Xander S.
·
Hey Barath C., thanks for trying out Phoenix and glad to hear the Q&A eval is working well for you. Can you elaborate on what you have in mind for UI-based evals? It sounds like maybe you are asking for a SaaS experience?
Barath C.
·
Hey Xander S. thanks for reaching out. I wan't to run these evals as part of some CI/CD pipelines. It would be nicer if there was an UI to visualise the results.
Jason
·
Barath C. The team has been talking through some ideas relative to this, expect us to eventually roll something out. Quick question - Are you most interested in saving/tracking Eval runs across datasets/dataframes or is the data you are evaluating related to LLM spans with LlamaIndex or LangChain? Maybe both?
Barath C.
·
Hey Jason , yes saving evaluations to potentially a database and able to visualise it with an UI. The data I am evaluating is the output of my own implementation I am not using LlamaIndex or LangChain for now.
👍1