Powered by
How to evaluate pre-generated outputs where input ...
# ❓|questions
r
rajendra.gola_86260
05/22/2025, 8:44 AM
I am able to evaluate the pre-generated LLM outputs using --model-outputs parameter. But in cases where input is required for evaluations along with pre-generated outputs it doesn't work. Please find attached the asserts.yaml and outputs.json file and also error in the terminal. Can someone please help here?
https://cdn.discordapp.com/attachments/1375031588252815391/1375031588521512970/image.png?ex=68303575&is=682ee3f5&hm=61a2be001d5ad26942d5efea26bbabe4f3fa56ad4b81a742e8d9640995fd5930&
https://cdn.discordapp.com/attachments/1375031588252815391/1375031588714446869/asserts.yaml?ex=68303575&is=682ee3f5&hm=091e358f20dfcf539923260958ad4a479d27f3ea9fe5a77898534381fc00edc2&
https://cdn.discordapp.com/attachments/1375031588252815391/1375031589066510336/outputs.json?ex=68303575&is=682ee3f5&hm=32aaab9c23d66fadb3f048949b4a0ce814a5dda39bb3ba34a1339a473250f10b&
Previous
Next