Table 2 Bayesian multilevel opinion changing model.
From: Testing the limits of large language models in debating humans
Conversation type | Odds ratio | Odds | CI (95%) |
|---|---|---|---|
hh | 1 | 0.12 | 0.068–0.185 |
ha | 0.16 | 0.02 | 0.003–0.046 |
ah | 2.99 | 0.36 | 0.203–0.574 |
aa | 1.97 | 0.23 | 0.162–0.319 |
Observations | 1284 | ||
|---|---|---|---|
\(N_{game}\) | 37 | ||
\(N_{player}\) | 185 | ||
\(\sigma ^2\) | 3.29 | ||
\(\tau _{00\ game}\) | 0.10 | ||
\(\tau _{00\ player}\) | 0.76 | ||
Marginal \(R^2\) | 0.030 | ||
Conditional \(R^2\) | 0.140 | ||
