Where did you find this claim? The first expression ${\bf e}_\sigma\cdot \partial_\mu{\bf e}_\nu$ is not covaraint. If one wrote instead ${\bf e}_\sigma\cdot \nabla_\mu{\bf e}_\nu$ then it makes sense because the Christoffel symbol is defined by
$$
\nabla_\mu {\bf e}_\nu = {\bf e}_\tau {\Gamma^\tau}_{\nu\mu}
$$
giving
$$
{\bf e}_\sigma\cdot \nabla_\mu{\bf e}_\nu= g_{\sigma\alpha} {\Gamma^\alpha}_{\nu\mu}
$$
and with $\nabla^\mu = g^{\mu\alpha}\nabla_\alpha $ and with the action of the covariant derivative on a covector being $\nabla_\alpha {\bf e}^\nu= - {\bf e}^{\tau}{\Gamma^\nu}_{\tau\mu}$ we get
$$
{\bf e}_\sigma\cdot \nabla^\mu{\bf e}^\nu = {\bf e}_\sigma (- {\bf e}^{\tau}){\Gamma^\nu}_{\tau\beta}g^{\beta\mu}=-{ \Gamma^\nu}_{\sigma
\beta}g^{\beta\mu}.
$$
So they differ by at least minus sign.
(sorry that I keep editing -- I keep making silly errors)