Speech Recognition/Whisper word level scores or confidence output

[wobbble]

Hey,
Big thanks for awesome project!

It possible to add score/confidence for word level output when using Speech Recognition/Whisper model?
Would appreciate any direction/comments or suggestion where to dig to add it.
Happy to submit PR if I will success in it.

Thanks!

[decoder-sh-david]

Seconded, I have also not been able to successfully get word-level timestamps while running on webgpu. Would love to have both!