I must wonder whether it's their watermarking initiative[1] forcing certain logit choices to produce watermarked text that ultimately causing the model to behave in a dumb manner.
From the little i understand that wouldnt be an issue because the model is ‘just’ using interchangeable words in a mathematical non-random way. Like using the same number of adjectives and the exct same words, but in a order that wouldn’t be mathematically plausible unless it was the watermark
From the little i understand that wouldnt be an issue because the model is ‘just’ using interchangeable words in a mathematical non-random way. Like using the same number of adjectives and the exct same words, but in a order that wouldn’t be mathematically plausible unless it was the watermark