PAIR-code / lit

The Learning Interpretability Tool: Interactively analyze ML models to understand their behavior in an extensible and framework agnostic interface.
https://pair-code.github.io/lit
Apache License 2.0
3.44k stars 350 forks source link

In the demo, only single-tower structures are shown. Are there examples of twin-tower structures, similar to Sentence-BERT? #1535

Open IcanDoItL opened 2 weeks ago

RyanMullins commented 2 weeks ago

This is probably a better question for @iftenney to answer, but I don't recall ever having a Sentence-BERT-like example in our demos.

Are you looking to analyze such a model with a specific UI module, or just generally explore how LIT could be used to evaluate such models?

IcanDoItL commented 2 weeks ago

I am looking to analyze such a model with a specific UI module. I hope the score is cosine similarity,not logits(is_regression) or softmax