4 ms·
Comparatively BARD models have relatively lower number of votes. I would wait until number of votes are in the same magnitude as other models.
by vinni2 3y ago
Comparatively BARD models have relatively lower number of votes. I would wait until number of votes are in the same magnitude as other models.
- danans 3y agoThat's why the site also shows the confidence interval.
- jsnell 3y ago3000 matches should be more than enough for an Elo-style ranking system to converge. But the notebook[0] linked to from the site includes a section with an equal number of matches being sampled for each pairwise comparison, and the results are pretty much the same. [0] https://colab.research.google.com/drive/1KdwokPjirkTmpO_P1WByFNFiqxWQquwH#scrollTo=5lBa9pdtzMA3 https://colab.research.google.com/drive/1KdwokPjirkTmpO_P1WB...