Puzzleheaded_Mall546

joined 10 months ago
 

Based on this image:

https://preview.redd.it/z5vf03e8r54c1.png?width=648&format=png&auto=webp&s=0a652e76ab2489135ed2327e8156029eacf274b7

Starling has better results than Zephyr DPO in all the metrics, Why ?

Shouldn't DPO be better than RLHF/RLAIF ?

[–] Puzzleheaded_Mall546@alien.top 1 points 10 months ago (1 children)

I am using the code of the space.

[–] Puzzleheaded_Mall546@alien.top 1 points 10 months ago (3 children)

I write text that is incomplete to see how it will translate it and the results is a coninuation of my text not the translation.

[–] Puzzleheaded_Mall546@alien.top 1 points 10 months ago (5 children)

I don't think its working.