diff --git a/R/Converter.R b/R/Converter.R
index a690a31..284cdf5 100644
--- a/R/Converter.R
+++ b/R/Converter.R
@@ -81,7 +81,8 @@ Converter <- R6Class("Converter",
     #' torch-converted model ([ConvertedModel]) is created and stored in
     #' the field `model`.
     #'
-    #' @param model ([`nn_sequential`], \code{\link[keras]{keras_model}},
+    #' @param model (\code{\link[torch]{nn_sequential}},
+    #' \code{\link[keras]{keras_model}},
     #' \code{\link[neuralnet]{neuralnet}} or `list`)\cr
     #' A trained neural network for classification or regression
     #' tasks to be interpreted. Only models from the following types or
diff --git a/R/DeepLift.R b/R/DeepLift.R
index 005d70e..242fc30 100644
--- a/R/DeepLift.R
+++ b/R/DeepLift.R
@@ -212,7 +212,8 @@ DeepSHAP <- R6Class(
     #' another variant is implemented, which treats a MaxPooling layer as an
     #' AveragePooling layer in the backward pass only, leading to an uniform
     #' distribution of the upper-layer contribution to the lower layer.\cr
-    #' @param data_ref ([`array`], [`data.frame`], [`torch_tensor`] or `list`)\cr
+    #' @param data_ref ([`array`], [`data.frame`],
+    #' \code{\link[torch]{torch_tensor}} or `list`)\cr
     #' The reference data which is used to estimate the conditional expectation.
     #' These must have the same format as the input data of the passed model to
     #' the converter object. This means either
diff --git a/R/GradienBased.R b/R/GradienBased.R
index 31a84cb..8f19c1b 100644
--- a/R/GradienBased.R
+++ b/R/GradienBased.R
@@ -319,7 +319,7 @@ IntegratedGradient <- R6Class(
     #' @param n (`integer(1)`)\cr
     #' Number of steps for the approximation of the integration path along
     #' \eqn{\alpha} (default: \eqn{50}).\cr
-    #' @param x_ref ([`array`], [`data.frame`], [`torch_tensor`] or `list`)\cr
+    #' @param x_ref ([`array`], [`data.frame`], \code{\link[torch]{torch_tensor}} or `list`)\cr
     #' The reference input for the IntegratedGradient method. This value
     #' must have the same format as the input data of the passed model to the
     #' converter object. This means either
@@ -723,7 +723,7 @@ ExpectedGradient <- R6Class(
     #' Number of samples from the distribution of reference values and number
     #' of samples for the approximation of the integration path along
     #' \eqn{\alpha} (default: \eqn{50}).\cr
-    #' @param data_ref ([`array`], [`data.frame`], [`torch_tensor`] or `list`)\cr
+    #' @param data_ref ([`array`], [`data.frame`], \code{\link[torch]{torch_tensor}} or `list`)\cr
     #' The reference inputs for the ExpectedGradient method. This value
     #' must have the same format as the input data of the passed model to the
     #' converter object. This means either
diff --git a/man/AgnosticWrapper.Rd b/man/AgnosticWrapper.Rd
index a55ef7c..a53df20 100644
--- a/man/AgnosticWrapper.Rd
+++ b/man/AgnosticWrapper.Rd
@@ -18,7 +18,7 @@ are available (all are wrapped by other packages):
 }
 }
 \section{Super class}{
-\code{innsight::InterpretingMethod} -> \code{AgnosticWrapper}
+\code{\link[innsight:InterpretingMethod]{innsight::InterpretingMethod}} -> \code{AgnosticWrapper}
 }
 \section{Public fields}{
 \if{html}{\out{<div class="r6-fields">}}
diff --git a/man/ConnectionWeights.Rd b/man/ConnectionWeights.Rd
index 508184f..775593f 100644
--- a/man/ConnectionWeights.Rd
+++ b/man/ConnectionWeights.Rd
@@ -156,7 +156,7 @@ Other methods:
 }
 \concept{methods}
 \section{Super class}{
-\code{innsight::InterpretingMethod} -> \code{ConnectionWeights}
+\code{\link[innsight:InterpretingMethod]{innsight::InterpretingMethod}} -> \code{ConnectionWeights}
 }
 \section{Public fields}{
 \if{html}{\out{<div class="r6-fields">}}
diff --git a/man/Converter.Rd b/man/Converter.Rd
index 5f977a3..4d8ca5a 100644
--- a/man/Converter.Rd
+++ b/man/Converter.Rd
@@ -275,7 +275,8 @@ the field \code{model}.
 \subsection{Arguments}{
 \if{html}{\out{<div class="arguments">}}
 \describe{
-\item{\code{model}}{(\code{\link{nn_sequential}}, \code{\link[keras]{keras_model}},
+\item{\code{model}}{(\code{\link[torch]{nn_sequential}},
+\code{\link[keras]{keras_model}},
 \code{\link[neuralnet]{neuralnet}} or \code{list})\cr
 A trained neural network for classification or regression
 tasks to be interpreted. Only models from the following types or
diff --git a/man/DeepLift.Rd b/man/DeepLift.Rd
index 1ec9213..b096dbf 100644
--- a/man/DeepLift.Rd
+++ b/man/DeepLift.Rd
@@ -157,7 +157,7 @@ Other methods:
 }
 \concept{methods}
 \section{Super class}{
-\code{innsight::InterpretingMethod} -> \code{DeepLift}
+\code{\link[innsight:InterpretingMethod]{innsight::InterpretingMethod}} -> \code{DeepLift}
 }
 \section{Public fields}{
 \if{html}{\out{<div class="r6-fields">}}
diff --git a/man/DeepSHAP.Rd b/man/DeepSHAP.Rd
index 331fc5f..b667c53 100644
--- a/man/DeepSHAP.Rd
+++ b/man/DeepSHAP.Rd
@@ -152,7 +152,7 @@ Other methods:
 }
 \concept{methods}
 \section{Super class}{
-\code{innsight::InterpretingMethod} -> \code{DeepSHAP}
+\code{\link[innsight:InterpretingMethod]{innsight::InterpretingMethod}} -> \code{DeepSHAP}
 }
 \section{Public fields}{
 \if{html}{\out{<div class="r6-fields">}}
@@ -300,7 +300,8 @@ often omitted.\cr}
 Name of the applied rule to calculate the
 contributions. Use either \code{'rescale'} or \code{'reveal_cancel'}. \cr}
 
-\item{\code{data_ref}}{(\code{\link{array}}, \code{\link{data.frame}}, \code{\link{torch_tensor}} or \code{list})\cr
+\item{\code{data_ref}}{(\code{\link{array}}, \code{\link{data.frame}},
+\code{\link[torch]{torch_tensor}} or \code{list})\cr
 The reference data which is used to estimate the conditional expectation.
 These must have the same format as the input data of the passed model to
 the converter object. This means either
diff --git a/man/ExpectedGradient.Rd b/man/ExpectedGradient.Rd
index f15c729..20f7ac1 100644
--- a/man/ExpectedGradient.Rd
+++ b/man/ExpectedGradient.Rd
@@ -155,7 +155,7 @@ Other methods:
 }
 \concept{methods}
 \section{Super classes}{
-\code{innsight::InterpretingMethod} -> \code{innsight::GradientBased} -> \code{ExpectedGradient}
+\code{\link[innsight:InterpretingMethod]{innsight::InterpretingMethod}} -> \code{\link[innsight:GradientBased]{innsight::GradientBased}} -> \code{ExpectedGradient}
 }
 \section{Public fields}{
 \if{html}{\out{<div class="r6-fields">}}
@@ -230,7 +230,7 @@ size \emph{(batch_size, dim_in)}, if e.g., the model has only one input layer, o
 upper point) for each of the input layers.\cr
 }}
 
-\item{\code{data_ref}}{(\code{\link{array}}, \code{\link{data.frame}}, \code{\link{torch_tensor}} or \code{list})\cr
+\item{\code{data_ref}}{(\code{\link{array}}, \code{\link{data.frame}}, \code{\link[torch]{torch_tensor}} or \code{list})\cr
 The reference inputs for the ExpectedGradient method. This value
 must have the same format as the input data of the passed model to the
 converter object. This means either
diff --git a/man/Gradient.Rd b/man/Gradient.Rd
index 41f7320..5e8fbab 100644
--- a/man/Gradient.Rd
+++ b/man/Gradient.Rd
@@ -154,7 +154,7 @@ Other methods:
 }
 \concept{methods}
 \section{Super classes}{
-\code{innsight::InterpretingMethod} -> \code{innsight::GradientBased} -> \code{Gradient}
+\code{\link[innsight:InterpretingMethod]{innsight::InterpretingMethod}} -> \code{\link[innsight:GradientBased]{innsight::GradientBased}} -> \code{Gradient}
 }
 \section{Methods}{
 \subsection{Public methods}{
diff --git a/man/GradientBased.Rd b/man/GradientBased.Rd
index 2204462..995e640 100644
--- a/man/GradientBased.Rd
+++ b/man/GradientBased.Rd
@@ -16,7 +16,7 @@ gradients w.r.t. to the input for given data. Implemented are:
 }
 }
 \section{Super class}{
-\code{innsight::InterpretingMethod} -> \code{GradientBased}
+\code{\link[innsight:InterpretingMethod]{innsight::InterpretingMethod}} -> \code{GradientBased}
 }
 \section{Public fields}{
 \if{html}{\out{<div class="r6-fields">}}
diff --git a/man/IntegratedGradient.Rd b/man/IntegratedGradient.Rd
index 662e957..322a6bb 100644
--- a/man/IntegratedGradient.Rd
+++ b/man/IntegratedGradient.Rd
@@ -159,7 +159,7 @@ Other methods:
 }
 \concept{methods}
 \section{Super classes}{
-\code{innsight::InterpretingMethod} -> \code{innsight::GradientBased} -> \code{IntegratedGradient}
+\code{\link[innsight:InterpretingMethod]{innsight::InterpretingMethod}} -> \code{\link[innsight:GradientBased]{innsight::GradientBased}} -> \code{IntegratedGradient}
 }
 \section{Public fields}{
 \if{html}{\out{<div class="r6-fields">}}
@@ -234,7 +234,7 @@ size \emph{(batch_size, dim_in)}, if e.g., the model has only one input layer, o
 upper point) for each of the input layers.\cr
 }}
 
-\item{\code{x_ref}}{(\code{\link{array}}, \code{\link{data.frame}}, \code{\link{torch_tensor}} or \code{list})\cr
+\item{\code{x_ref}}{(\code{\link{array}}, \code{\link{data.frame}}, \code{\link[torch]{torch_tensor}} or \code{list})\cr
 The reference input for the IntegratedGradient method. This value
 must have the same format as the input data of the passed model to the
 converter object. This means either
diff --git a/man/LIME.Rd b/man/LIME.Rd
index 02e64c9..3d912ae 100644
--- a/man/LIME.Rd
+++ b/man/LIME.Rd
@@ -131,7 +131,7 @@ Other methods:
 }
 \concept{methods}
 \section{Super classes}{
-\code{innsight::InterpretingMethod} -> \code{innsight::AgnosticWrapper} -> \code{LIME}
+\code{\link[innsight:InterpretingMethod]{innsight::InterpretingMethod}} -> \code{\link[innsight:AgnosticWrapper]{innsight::AgnosticWrapper}} -> \code{LIME}
 }
 \section{Methods}{
 \subsection{Public methods}{
diff --git a/man/LRP.Rd b/man/LRP.Rd
index 0dd6f45..f7f9093 100644
--- a/man/LRP.Rd
+++ b/man/LRP.Rd
@@ -170,7 +170,7 @@ Other methods:
 }
 \concept{methods}
 \section{Super class}{
-\code{innsight::InterpretingMethod} -> \code{LRP}
+\code{\link[innsight:InterpretingMethod]{innsight::InterpretingMethod}} -> \code{LRP}
 }
 \section{Public fields}{
 \if{html}{\out{<div class="r6-fields">}}
diff --git a/man/SHAP.Rd b/man/SHAP.Rd
index 8428e1c..a5cd2b3 100644
--- a/man/SHAP.Rd
+++ b/man/SHAP.Rd
@@ -126,7 +126,7 @@ Other methods:
 }
 \concept{methods}
 \section{Super classes}{
-\code{innsight::InterpretingMethod} -> \code{innsight::AgnosticWrapper} -> \code{SHAP}
+\code{\link[innsight:InterpretingMethod]{innsight::InterpretingMethod}} -> \code{\link[innsight:AgnosticWrapper]{innsight::AgnosticWrapper}} -> \code{SHAP}
 }
 \section{Methods}{
 \subsection{Public methods}{
diff --git a/man/SmoothGrad.Rd b/man/SmoothGrad.Rd
index 62b33f9..8ca6d48 100644
--- a/man/SmoothGrad.Rd
+++ b/man/SmoothGrad.Rd
@@ -150,7 +150,7 @@ Other methods:
 }
 \concept{methods}
 \section{Super classes}{
-\code{innsight::InterpretingMethod} -> \code{innsight::GradientBased} -> \code{SmoothGrad}
+\code{\link[innsight:InterpretingMethod]{innsight::InterpretingMethod}} -> \code{\link[innsight:GradientBased]{innsight::GradientBased}} -> \code{SmoothGrad}
 }
 \section{Public fields}{
 \if{html}{\out{<div class="r6-fields">}}