Difference Between run_evals and llm_classify in Arize

·Mar 28, 2024 07:57 AM

Hi Arize Team, can you briefly explain what the difference is between run_evals (referred to in the Quickstart) and llm_classify (referred to in the How-to) and when to use which approach? Thank you!

✅1

2 comments

· Sorted by Oldest

Phoenix Support

Difference Between run_evals and llm_classify in Arize

Daniel

·Mar 28, 2024 07:57 AM

Hi Arize Team, can you briefly explain what the difference is between run_evals (referred to in the Quickstart) and llm_classify (referred to in the How-to) and when to use which approach? Thank you!

✅1

2 comments

· Sorted by Oldest

Mikyo
·
Hey Daniel our classification based evaluators (which are all of them) use classify under the hood and then assign a score based on that classification. So in essence they are the same. If you want to run multiple evals in one go, the evaluators are designed for parallel execution while LLM classify is if you need to build custom evaluators. Hope that makes sense!
Daniel
·
Makes perfectly sense! As always your support is priceless - thank you!
1

Mikyo
·
Hey Daniel our classification based evaluators (which are all of them) use classify under the hood and then assign a score based on that classification. So in essence they are the same. If you want to run multiple evals in one go, the evaluators are designed for parallel execution while LLM classify is if you need to build custom evaluators. Hope that makes sense!
Daniel
·
Makes perfectly sense! As always your support is priceless - thank you!
1