stanford-futuredata / ARES

Automated Evaluation of RAG Systems
https://ares-ai.vercel.app/
Apache License 2.0
486 stars 53 forks source link

What is meant by contrastive learning objective? #74

Open majdsalman opened 1 month ago

majdsalman commented 1 month ago

Having read the paper and the code I don't understand what is meant by "These judges are lightweight models fine-tuned against a contrastive learning objective.".

The loss function is a cross entropy loss and the task is a classification task (yes/no) for belonging or not belonging in a class. There is no notion in the code of maximizing the separation of these classes in the embeddings space. The phrasing adds some confusion since contrastive typically means something else.