4 ms·
Fair; have tried to combat this issue in a few ways. Each model's position is scored against outside political-science data (Chapel Hill Expert Survey for part
by mektrik 3mo ago
Fair; have tried to combat this issue in a few ways.
Each model's position is scored against outside political-science data (Chapel Hill Expert Survey for party positions, World Values Survey for where populations sit).
The stance coding is done by a separate model with a published prompt + a second model from a different lab re-scores a sample and we publish where the two disagree.
So not perfect but (as far as I can tell) one of the more defensible approaches.
- riddley 3mo agoIsn't it completely subjective to the point in time Overton window?
- rappatic 3mo agoYes, there's no way to avoid this really. You have to pick some point on the graph as the origin.
- ifwinterco 3mo agoI’d argue the entire concept of being “unbiased” is flawed. All humans are biased all of the time, bias towards “the centre” is still a form of bias. One cannot escape ideology, one merely swaps one ideology for another