I am using a gamma response in glmer() from lme4. Some rows in my data set represents averages from n_i observations, while most observations just are individual measurements. My idea was to incorporate n_i as weigths in lme4's glmer(). Specifically, I am using a gamma response with logarithmic link function. The question is, what are the weights doing?

The glmer documentation only says "an optional vector of ‘prior weights’ to be used in the fitting process. Should be NULL or a numeric vector."

Should I interpret that as that they just multiplies in the likelihood, so that it is roughly like saying that if n_i=10 that observation gets 10 times the influence in the likelihood?

I assume it is not like weights in weighted least squares (WLS) where the weights are meant to account for unequal variances?