diff --git a/man/Lrnr_haldensify.Rd b/man/Lrnr_haldensify.Rd
index 2358e816..5b636244 100644
--- a/man/Lrnr_haldensify.Rd
+++ b/man/Lrnr_haldensify.Rd
@@ -33,6 +33,15 @@ to be divided into.
 sequence of values of the regulariztion parameter of the Lasso regression,
 to be passed to to \code{\link[hal9001]{fit_hal}}.
 }
+\item{\code{trim_dens = 1/sqrt(n)}}{A \code{numeric} giving the minimum
+allowed value of the resultant density predictions. Any predicted
+density values below this tolerance threshold are set to the indicated
+minimum. The default is to use the inverse of the square root of the
+sample size of the prediction set, i.e., 1/sqrt(n); another notable
+choice is 1/sqrt(n)/log(n). If there are observations in the prediction
+set with values of \code{new_A} outside of the support of the training
+set, their predictions are similarly truncated.
+}
 \item{\code{...}}{ Other parameters passed directly to
 \code{\link[haldensify]{haldensify}}. See its documentation for details.
 }
diff --git a/tests/testthat/test-lightgbm.R b/tests/testthat/test-lightgbm.R
index 5c7a91e9..9493c1e6 100644
--- a/tests/testthat/test-lightgbm.R
+++ b/tests/testthat/test-lightgbm.R
@@ -24,6 +24,7 @@ test_learner <- function(learner, task, ...) {
   # test learner prediction
   test_that("Learner can generate training set predictions", {
     skip_on_os("windows")
+    fit_obj <- learner_obj$train(task)
     train_preds <- fit_obj$predict()
     expect_equal(
       sl3:::safe_dim(train_preds)[1], length(task$Y)
@@ -33,11 +34,15 @@ test_learner <- function(learner, task, ...) {
   # test learner chaining
   test_that("Chaining returns a task", {
     skip_on_os("windows")
+    fit_obj <- learner_obj$train(task)
     chained_task <- fit_obj$chain()
     expect_true(is(chained_task, "sl3_Task"))
   })
+
   test_that("Chaining returns the correct number of rows", {
     skip_on_os("windows")
+    fit_obj <- learner_obj$train(task)
+    chained_task <- fit_obj$chain()
     expect_equal(nrow(chained_task$X), nrow(task$X))
   })
 }