diff --git a/R/class-kap.R b/R/class-kap.R
index 4566530d..2519da75 100644
--- a/R/class-kap.R
+++ b/R/class-kap.R
@@ -178,7 +178,7 @@ make_weighting_matrix <- function(weighting, n_levels, call = caller_env()) {
 
 validate_weighting <- function(x, call = caller_env()) {
   if (!is_string(x)) {
-    abort("`weighting` must be a string.", call = call)
+    cli::cli_abort("{.arg weighting} must be a string.", call = call)
   }
 
   ok <- is_no_weighting(x) ||
@@ -186,7 +186,11 @@ validate_weighting <- function(x, call = caller_env()) {
     is_quadratic_weighting(x)
 
   if (!ok) {
-    abort("`weighting` must be 'none', 'linear', or 'quadratic'.", call = call)
+    cli::cli_abort(
+      "{.arg weighting} must be {.val none}, {.val linear}, or \\
+      {.val quadratic}, not {.val {x}}.",
+      call = call
+    )
   }
 
   invisible(x)
diff --git a/R/class-mcc.R b/R/class-mcc.R
index 72808f5b..a3461f52 100644
--- a/R/class-mcc.R
+++ b/R/class-mcc.R
@@ -161,7 +161,10 @@ mcc_multiclass_impl <- function(C) {
 
 check_mcc_data <- function(data) {
   if (!is.double(data) && !is.matrix(data)) {
-    abort("`data` should be a double matrix at this point.", .internal = TRUE)
+    cli::cli_abort(
+      "{.arg data} should be a double matrix at this point.",
+      .internal = TRUE
+    )
   }
   invisible()
 }
diff --git a/R/conf_mat.R b/R/conf_mat.R
index 5a384679..d6ff1b3b 100644
--- a/R/conf_mat.R
+++ b/R/conf_mat.R
@@ -223,7 +223,10 @@ conf_mat_impl <- function(truth, estimate, case_weights, call = caller_env()) {
   check_class_metric(truth, estimate, case_weights, estimator, call = call)
 
   if (length(levels(truth)) < 2) {
-    abort("`truth` must have at least 2 factor levels.", call = call)
+    cli::cli_abort(
+      "{.arg truth} must have at least 2 factor levels.",
+      call = call
+    )
   }
 
   yardstick_table(
@@ -245,7 +248,9 @@ conf_mat.table <- function(data, ...) {
   num_lev <- length(class_lev)
 
   if (num_lev < 2) {
-    abort("There must be at least 2 factors levels in the `data`")
+    cli::cli_abort(
+      "There must be at least 2 factors levels in the {.arg data}."
+    )
   }
 
   structure(
diff --git a/R/misc.R b/R/misc.R
index 953391e2..13f74b3b 100644
--- a/R/misc.R
+++ b/R/misc.R
@@ -4,7 +4,7 @@
 
 pos_val <- function(xtab, event_level) {
   if (!all(dim(xtab) == 2)) {
-    abort("Only relevant for 2x2 tables")
+    cli::cli_abort("Only relevant for 2x2 tables.")
   }
 
   if (is_event_first(event_level)) {
@@ -16,7 +16,7 @@ pos_val <- function(xtab, event_level) {
 
 neg_val <- function(xtab, event_level) {
   if (!all(dim(xtab) == 2)) {
-    abort("Only relevant for 2x2 tables")
+    cli::cli_abort("Only relevant for 2x2 tables.")
   }
 
   if (is_event_first(event_level)) {
@@ -67,19 +67,19 @@ as_factor_from_class_pred <- function(x) {
   }
 
   if (!is_installed("probably")) {
-    abort(paste0(
-      "A <class_pred> input was detected, but the probably package ",
-      "isn't installed. Install probably to be able to convert <class_pred> ",
-      "to <factor>."
-    ))
+    cli::cli_abort(
+      "A {.cls class_pred} input was detected, but the {.pkg probably} \\
+      package isn't installed. Install {.pkg probably} to be able to convert \\
+      {.cls class_pred} to {.cls factor}."
+    )
   }
   probably::as.factor(x)
 }
 
 abort_if_class_pred <- function(x, call = caller_env()) {
   if (is_class_pred(x)) {
-    abort(
-      "`truth` should not a `class_pred` object.",
+    cli::cli_abort(
+      "{.arg truth} should not a {.cls class_pred} object.",
       call = call
     )
   }
@@ -186,10 +186,18 @@ yardstick_cov <- function(truth,
 
   size <- vec_size(truth)
   if (size != vec_size(estimate)) {
-    abort("`truth` and `estimate` must be the same size.", .internal = TRUE)
+    cli::cli_abort(
+      "{.arg truth} ({vec_size(truth)}) and \\
+      {.arg estimate} ({vec_size(estimate)}) must be the same size.",
+      .internal = TRUE
+    )
   }
   if (size != vec_size(case_weights)) {
-    abort("`truth` and `case_weights` must be the same size.", .internal = TRUE)
+    cli::cli_abort(
+      "{.arg truth} ({vec_size(truth)}) and \\
+      {.arg case_weights} ({vec_size(case_weights)}) must be the same size.",
+      .internal = TRUE
+    )
   }
 
   if (size == 0L || size == 1L) {
@@ -232,10 +240,18 @@ yardstick_cor <- function(truth,
 
   size <- vec_size(truth)
   if (size != vec_size(estimate)) {
-    abort("`truth` and `estimate` must be the same size.", .internal = TRUE)
+    cli::cli_abort(
+      "{.arg truth} ({vec_size(truth)}) and \\
+      {.arg estimate} ({vec_size(estimate)}) must be the same size.",
+      .internal = TRUE
+    )
   }
   if (size != vec_size(case_weights)) {
-    abort("`truth` and `case_weights` must be the same size.", .internal = TRUE)
+    cli::cli_abort(
+      "{.arg truth} ({vec_size(truth)}) and \\
+      {.arg case_weights} ({vec_size(case_weights)}) must be the same size.",
+      .internal = TRUE
+    )
   }
 
   if (size == 0L || size == 1L) {
@@ -345,14 +361,17 @@ weighted_quantile <- function(x, weights, probabilities) {
 
   size <- vec_size(x)
   if (size != vec_size(weights)) {
-    abort("`x` and `weights` must have the same size.")
+    cli::cli_abort(
+      "{.arg x} ({vec_size(x)}) and {.arg weights} ({vec_size(weights)}) \\
+      must have the same size."
+    )
   }
 
   if (any(is.na(probabilities))) {
-    abort("`probabilities` can't be missing.")
+    cli::cli_abort("{.arg probabilities} can't have missing values.")
   }
   if (any(probabilities > 1 | probabilities < 0)) {
-    abort("`probabilities` must be within `[0, 1]`.")
+    cli::cli_abort("{.arg probabilities} must be within `[0, 1]`.")
   }
 
   if (size == 0L) {
@@ -397,20 +416,33 @@ yardstick_table <- function(truth, estimate, ..., case_weights = NULL) {
   }
 
   if (!is.factor(truth)) {
-    abort("`truth` must be a factor.", .internal = TRUE)
+    cli::cli_abort(
+      "{.arg truth} must be a factor, not {.obj_type_friendly {truth}}.",
+      .internal = TRUE
+    )
   }
   if (!is.factor(estimate)) {
-    abort("`estimate` must be a factor.", .internal = TRUE)
+    cli::cli_abort(
+      "{.arg estimate} must be a factor, not {.obj_type_friendly {estimate}}.",
+      .internal = TRUE
+    )
   }
 
   levels <- levels(truth)
   n_levels <- length(levels)
 
   if (!identical(levels, levels(estimate))) {
-    abort("`truth` and `estimate` must have the same levels in the same order.", .internal = TRUE)
+    cli::cli_abort(
+      "{.arg truth} and {.arg estimate} must have the same levels in the same \\
+      order.",
+      .internal = TRUE
+    )
   }
   if (n_levels < 2) {
-    abort("`truth` must have at least 2 factor levels.", .internal = TRUE)
+    cli::cli_abort(
+      "{.arg truth} must have at least 2 factor levels.",
+      .internal = TRUE
+    )
   }
 
   # Supply `estimate` first to get it to correspond to the row names.
@@ -447,14 +479,17 @@ yardstick_truth_table <- function(truth, ..., case_weights = NULL) {
   abort_if_class_pred(truth)
 
   if (!is.factor(truth)) {
-    abort("`truth` must be a factor.", .internal = TRUE)
+    cli::cli_abort("{.arg truth} must be a factor.", .internal = TRUE)
   }
 
   levels <- levels(truth)
   n_levels <- length(levels)
 
   if (n_levels < 2) {
-    abort("`truth` must have at least 2 factor levels.", .internal = TRUE)
+    cli::cli_abort(
+      "{.arg truth} must have at least 2 factor levels.",
+      .internal = TRUE
+    )
   }
 
   # Always return a double matrix for type stability
diff --git a/R/num-huber_loss.R b/R/num-huber_loss.R
index a31eeb11..d418c064 100644
--- a/R/num-huber_loss.R
+++ b/R/num-huber_loss.R
@@ -85,12 +85,7 @@ huber_loss_impl <- function(truth,
   # Weighted Huber Loss implementation confirmed against matlab:
   # https://www.mathworks.com/help/deeplearning/ref/dlarray.huber.html
 
-  if (!is_bare_numeric(delta, n = 1L)) {
-    abort("`delta` must be a single numeric value.", call = call)
-  }
-  if (!(delta >= 0)) {
-    abort("`delta` must be a positive value.", call = call)
-  }
+  check_number_decimal(delta, min = 0, call = call)
 
   a <- truth - estimate
   abs_a <- abs(a)
diff --git a/R/num-mase.R b/R/num-mase.R
index bde04db7..378b76ff 100644
--- a/R/num-mase.R
+++ b/R/num-mase.R
@@ -114,8 +114,8 @@ mase_impl <- function(truth,
                       mae_train = NULL,
                       case_weights = NULL,
                       call = caller_env()) {
-  validate_m(m, call = call)
-  validate_mae_train(mae_train, call = call)
+  check_number_whole(m, min = 0, call = call)
+  check_number_decimal(mae_train, min = 0, allow_null = TRUE, call = call)
 
   if (is.null(mae_train)) {
     validate_truth_m(truth, m, call = call)
@@ -139,46 +139,12 @@ mase_impl <- function(truth,
   out
 }
 
-validate_m <- function(m, call = caller_env()) {
-  abort_msg <- "`m` must be a single positive integer value."
-
-  if (!is_integerish(m, n = 1L)) {
-    abort(abort_msg, call = call)
-  }
-
-  if (!(m > 0)) {
-    abort(abort_msg, call = call)
-  }
-
-  invisible(m)
-}
-
-validate_mae_train <- function(mae_train, call = caller_env()) {
-  if (is.null(mae_train)) {
-    return(invisible(mae_train))
-  }
-
-  is_single_numeric <- is_bare_numeric(mae_train, n = 1L)
-  abort_msg <- "`mae_train` must be a single positive numeric value."
-
-  if (!is_single_numeric) {
-    abort(abort_msg, call = call)
-  }
-
-  if (!(mae_train > 0)) {
-    abort(abort_msg, call = call)
-  }
-
-  invisible(mae_train)
-}
-
 validate_truth_m <- function(truth, m, call = caller_env()) {
   if (length(truth) <= m) {
-    abort(paste0(
-      "`truth` must have a length greater than `m` ",
-      "to compute the out-of-sample naive mean absolute error."
-    ), call = call)
+    cli::cli_abort(
+      "{.arg truth} ({length(truth)}) must have a length greater than \\
+      {.arg m} ({m}) to compute the out-of-sample naive mean absolute error.",
+      call = call
+    )
   }
-
-  invisible(truth)
 }
diff --git a/R/num-pseudo_huber_loss.R b/R/num-pseudo_huber_loss.R
index ac616a29..b6724d62 100644
--- a/R/num-pseudo_huber_loss.R
+++ b/R/num-pseudo_huber_loss.R
@@ -86,12 +86,7 @@ huber_loss_pseudo_impl <- function(truth,
                                    delta,
                                    case_weights,
                                    call = caller_env()) {
-  if (!is_bare_numeric(delta, n = 1L)) {
-    abort("`delta` must be a single numeric value.", call = call)
-  }
-  if (!(delta >= 0)) {
-    abort("`delta` must be a positive value.", call = call)
-  }
+  check_number_decimal(delta, min = 0, call = call)
 
   a <- truth - estimate
   loss <- delta^2 * (sqrt(1 + (a / delta)^2) - 1)
diff --git a/R/prob-binary-thresholds.R b/R/prob-binary-thresholds.R
index e38a15cb..9f89074c 100644
--- a/R/prob-binary-thresholds.R
+++ b/R/prob-binary-thresholds.R
@@ -16,19 +16,36 @@ binary_threshold_curve <- function(truth,
   case_weights <- vec_cast(case_weights, to = double())
 
   if (!is.factor(truth)) {
-    abort("`truth` must be a factor.", .internal = TRUE)
+    cli::cli_abort(
+      "{.arg truth} must be a factor, not {.obj_friendly_type {truth}}.",
+      .internal = TRUE
+    )
   }
   if (length(levels(truth)) != 2L) {
-    abort("`truth` must have two levels.", .internal = TRUE)
+    cli::cli_abort(
+      "{.arg truth} must have two levels, not {length(levels(truth))}.",
+      .internal = TRUE
+    )
   }
   if (!is.numeric(estimate)) {
-    abort("`estimate` must be numeric.", .internal = TRUE)
+    cli::cli_abort(
+      "{.arg estimate} must be numeric, not {.obj_friendly_type {estimate}}.",
+      .internal = TRUE
+    )
   }
   if (length(truth) != length(estimate)) {
-    abort("`truth` and `estimate` must be the same length.", .internal = TRUE)
+    cli::cli_abort(
+      "{.arg truth} ({length(truth)}) and \\
+      {.arg estimate} ({length(estimate)}) must be the same length.",
+      .internal = TRUE
+    )
   }
   if (length(truth) != length(case_weights)) {
-    abort("`truth` and `case_weights` must be the same length.", .internal = TRUE)
+    cli::cli_abort(
+      "{.arg truth} ({length(truth)}) and \\
+      {.arg case_weights} ({length(case_weights)}) must be the same length.",
+      .internal = TRUE
+    )
   }
 
   truth <- unclass(truth)
diff --git a/R/prob-classification_cost.R b/R/prob-classification_cost.R
index 72b49188..7042e83b 100644
--- a/R/prob-classification_cost.R
+++ b/R/prob-classification_cost.R
@@ -221,48 +221,60 @@ validate_costs <- function(costs, levels) {
     return(costs)
   }
 
-  if (!is.data.frame(costs)) {
-    abort("`costs` must be `NULL` or a data.frame.")
-  }
+  check_data_frame(costs, allow_null = TRUE)
 
   columns <- names(costs)
   if (length(columns) != 3L) {
-    abort("`costs` must be a data.frame with 3 columns.")
+    cli::cli_abort(
+      "{.arg costs} must be a data.frame with 3 columns, not {length(columns)}."
+    )
   }
 
   ok <- identical(sort(columns), sort(c("truth", "estimate", "cost")))
   if (!ok) {
-    abort("`costs` must have columns: 'truth', 'estimate', and 'cost'.")
+    cli::cli_abort(
+      "{.arg costs} must have columns: \\
+      {.val truth}, {.val estimate}, and {.val cost}. Not {columns}."
+    )
   }
 
   if (is.factor(costs$truth)) {
     costs$truth <- as.character(costs$truth)
   }
   if (!is.character(costs$truth)) {
-    abort("`costs$truth` must be a character or factor column.")
+    cli::cli_abort(
+      "{.code costs$truth} must be a character or factor column, \\
+      not {.obj_friendly_type {costs$truth}}."
+    )
   }
 
   if (is.factor(costs$estimate)) {
     costs$estimate <- as.character(costs$estimate)
   }
   if (!is.character(costs$estimate)) {
-    abort("`costs$estimate` must be a character or factor column.")
+    cli::cli_abort(
+      "{.code costs$estimate} must be a character or factor column, \\
+      not {.obj_friendly_type {costs$estimate}}."
+      )
   }
 
   if (!is.numeric(costs$cost)) {
-    abort("`costs$cost` must be a numeric column.")
+    cli::cli_abort(
+      "{.code costs$cost} must be a numeric column, \\
+      not {.obj_friendly_type {costs$cost}}."
+    )
   }
 
   ok <- all(costs$truth %in% levels)
   if (is_false(ok)) {
     levels <- quote_and_collapse(levels)
-    abort(paste0("`costs$truth` can only contain ", levels, "."))
+    cli::cli_abort("{.code costs$truth} can only contain {levels}.")
   }
 
   ok <- all(costs$estimate %in% levels)
   if (is_false(ok)) {
     levels <- quote_and_collapse(levels)
-    abort(paste0("`costs$estimate` can only contain ", levels, "."))
+    cli::cli_abort("{.code costs$estimate} can only contain {levels}.")
   }
 
   pairs <- costs[c("truth", "estimate")]
@@ -270,7 +282,10 @@ validate_costs <- function(costs, levels) {
 
   not_ok <- vec_duplicate_any(pairs)
   if (not_ok) {
-    abort("`costs` cannot have duplicate 'truth' / 'estimate' combinations.")
+    cli::cli_abort(
+      "{.field costs} cannot have duplicate \\
+      {.field truth} / {.field estimate} combinations."
+    )
   }
 
   out <- generate_all_level_combinations(levels)
diff --git a/R/prob-roc_auc.R b/R/prob-roc_auc.R
index 97f1b202..728dd64d 100644
--- a/R/prob-roc_auc.R
+++ b/R/prob-roc_auc.R
@@ -163,7 +163,10 @@ roc_auc_estimator_impl <- function(truth,
     roc_auc_binary(truth, estimate, event_level, case_weights)
   } else if (estimator == "hand_till") {
     if (!is.null(case_weights)) {
-      abort("`case_weights` should be `NULL` at this point for hand-till.", .internal = TRUE)
+      cli::cli_abort(
+        "{.arg case_weights} should be `NULL` at this point for hand-till.",
+        .internal = TRUE
+      )
     }
 
     roc_auc_hand_till(truth, estimate)
@@ -258,7 +261,10 @@ finalize_estimator_roc_auc <- function(x,
       estimator <- "macro"
     } else {
       # Manually chose hand-till and specified case weights. Not compatible!
-      abort("Can't specify both `estimator = 'hand_till'` and `case_weights`.")
+      cli::cli_abort(
+        "Can't specify both {.code estimator = 'hand_till'} and \\
+        {.code case_weights}."
+      )
     }
   }
 
@@ -396,7 +402,7 @@ warn_roc_truth_no_control <- function(control) {
   )
 }
 stop_roc_truth_no_control <- function(control) {
-  abort(
+  cli::cli_abort(
     msg_roc_truth_no_control(control),
     class = "yardstick_error_roc_truth_no_control"
   )
@@ -415,7 +421,7 @@ warn_roc_truth_no_event <- function(event) {
   )
 }
 stop_roc_truth_no_event <- function(event) {
-  abort(
+  cli::cli_abort(
     msg_roc_truth_no_event(event),
     class = "yardstick_error_roc_truth_no_event"
   )
diff --git a/tests/testthat/_snaps/class-accuracy.md b/tests/testthat/_snaps/class-accuracy.md
index 0ebfd5d5..61aceec8 100644
--- a/tests/testthat/_snaps/class-accuracy.md
+++ b/tests/testthat/_snaps/class-accuracy.md
@@ -4,5 +4,5 @@
       accuracy_vec(cp_truth, cp_estimate)
     Condition
       Error in `accuracy_vec()`:
-      ! `truth` should not a `class_pred` object.
+      ! `truth` should not a <class_pred> object.
 
diff --git a/tests/testthat/_snaps/class-detection_prevalence.md b/tests/testthat/_snaps/class-detection_prevalence.md
index 2f2145f0..629917e8 100644
--- a/tests/testthat/_snaps/class-detection_prevalence.md
+++ b/tests/testthat/_snaps/class-detection_prevalence.md
@@ -4,5 +4,5 @@
       detection_prevalence_vec(cp_truth, cp_estimate)
     Condition
       Error in `detection_prevalence_vec()`:
-      ! `truth` should not a `class_pred` object.
+      ! `truth` should not a <class_pred> object.
 
diff --git a/tests/testthat/_snaps/class-f_meas.md b/tests/testthat/_snaps/class-f_meas.md
index 08132b45..ae33112d 100644
--- a/tests/testthat/_snaps/class-f_meas.md
+++ b/tests/testthat/_snaps/class-f_meas.md
@@ -80,5 +80,5 @@
       f_meas_vec(cp_truth, cp_estimate)
     Condition
       Error in `f_meas_vec()`:
-      ! `truth` should not a `class_pred` object.
+      ! `truth` should not a <class_pred> object.
 
diff --git a/tests/testthat/_snaps/class-j_index.md b/tests/testthat/_snaps/class-j_index.md
index 8463d479..aa868a93 100644
--- a/tests/testthat/_snaps/class-j_index.md
+++ b/tests/testthat/_snaps/class-j_index.md
@@ -52,5 +52,5 @@
       j_index_vec(cp_truth, cp_estimate)
     Condition
       Error in `j_index_vec()`:
-      ! `truth` should not a `class_pred` object.
+      ! `truth` should not a <class_pred> object.
 
diff --git a/tests/testthat/_snaps/class-kap.md b/tests/testthat/_snaps/class-kap.md
index bd2338ef..e9934ee0 100644
--- a/tests/testthat/_snaps/class-kap.md
+++ b/tests/testthat/_snaps/class-kap.md
@@ -12,7 +12,7 @@
       kap(three_class, truth = "obs", estimate = "pred", weighting = "not right")
     Condition
       Error in `kap()`:
-      ! `weighting` must be 'none', 'linear', or 'quadratic'.
+      ! `weighting` must be "none", "linear", or "quadratic", not "not right".
 
 # work with class_pred input
 
@@ -20,5 +20,5 @@
       kap_vec(cp_truth, cp_estimate)
     Condition
       Error in `kap_vec()`:
-      ! `truth` should not a `class_pred` object.
+      ! `truth` should not a <class_pred> object.
 
diff --git a/tests/testthat/_snaps/class-mcc.md b/tests/testthat/_snaps/class-mcc.md
index e59805a5..1e046607 100644
--- a/tests/testthat/_snaps/class-mcc.md
+++ b/tests/testthat/_snaps/class-mcc.md
@@ -4,5 +4,5 @@
       mcc_vec(cp_truth, cp_estimate)
     Condition
       Error in `mcc_vec()`:
-      ! `truth` should not a `class_pred` object.
+      ! `truth` should not a <class_pred> object.
 
diff --git a/tests/testthat/_snaps/class-npv.md b/tests/testthat/_snaps/class-npv.md
index 0ebfd5d5..61aceec8 100644
--- a/tests/testthat/_snaps/class-npv.md
+++ b/tests/testthat/_snaps/class-npv.md
@@ -4,5 +4,5 @@
       accuracy_vec(cp_truth, cp_estimate)
     Condition
       Error in `accuracy_vec()`:
-      ! `truth` should not a `class_pred` object.
+      ! `truth` should not a <class_pred> object.
 
diff --git a/tests/testthat/_snaps/class-ppv.md b/tests/testthat/_snaps/class-ppv.md
index 29fa28d8..c0332ae5 100644
--- a/tests/testthat/_snaps/class-ppv.md
+++ b/tests/testthat/_snaps/class-ppv.md
@@ -14,5 +14,5 @@
       ppv_vec(cp_truth, cp_estimate)
     Condition
       Error in `ppv_vec()`:
-      ! `truth` should not a `class_pred` object.
+      ! `truth` should not a <class_pred> object.
 
diff --git a/tests/testthat/_snaps/class-precision.md b/tests/testthat/_snaps/class-precision.md
index 8d262769..4fa54bda 100644
--- a/tests/testthat/_snaps/class-precision.md
+++ b/tests/testthat/_snaps/class-precision.md
@@ -38,5 +38,5 @@
       precision_vec(cp_truth, cp_estimate)
     Condition
       Error in `precision_vec()`:
-      ! `truth` should not a `class_pred` object.
+      ! `truth` should not a <class_pred> object.
 
diff --git a/tests/testthat/_snaps/class-recall.md b/tests/testthat/_snaps/class-recall.md
index dd21bb27..6267250c 100644
--- a/tests/testthat/_snaps/class-recall.md
+++ b/tests/testthat/_snaps/class-recall.md
@@ -26,5 +26,5 @@
       recall_vec(cp_truth, cp_estimate)
     Condition
       Error in `recall_vec()`:
-      ! `truth` should not a `class_pred` object.
+      ! `truth` should not a <class_pred> object.
 
diff --git a/tests/testthat/_snaps/class-sens.md b/tests/testthat/_snaps/class-sens.md
index 987e1f31..71ea7667 100644
--- a/tests/testthat/_snaps/class-sens.md
+++ b/tests/testthat/_snaps/class-sens.md
@@ -26,5 +26,5 @@
       sensitivity_vec(cp_truth, cp_estimate)
     Condition
       Error in `sensitivity_vec()`:
-      ! `truth` should not a `class_pred` object.
+      ! `truth` should not a <class_pred> object.
 
diff --git a/tests/testthat/_snaps/class-spec.md b/tests/testthat/_snaps/class-spec.md
index ec09ccb7..cd67e613 100644
--- a/tests/testthat/_snaps/class-spec.md
+++ b/tests/testthat/_snaps/class-spec.md
@@ -25,5 +25,5 @@
       specificity_vec(cp_truth, cp_estimate)
     Condition
       Error in `specificity_vec()`:
-      ! `truth` should not a `class_pred` object.
+      ! `truth` should not a <class_pred> object.
 
diff --git a/tests/testthat/_snaps/misc.md b/tests/testthat/_snaps/misc.md
index 56d1f961..b1d4e720 100644
--- a/tests/testthat/_snaps/misc.md
+++ b/tests/testthat/_snaps/misc.md
@@ -4,7 +4,7 @@
       yardstick_table(1, x)
     Condition
       Error in `yardstick_table()`:
-      ! `truth` must be a factor.
+      ! `truth` must be a factor, not a number.
       i This is an internal error that was detected in the yardstick package.
         Please report it at <https://github.com/tidymodels/yardstick/issues> with a reprex (<https://tidyverse.org/help/>) and the full backtrace.
 
@@ -14,7 +14,7 @@
       yardstick_table(x, 2)
     Condition
       Error in `yardstick_table()`:
-      ! `estimate` must be a factor.
+      ! `estimate` must be a factor, not a number.
       i This is an internal error that was detected in the yardstick package.
         Please report it at <https://github.com/tidymodels/yardstick/issues> with a reprex (<https://tidyverse.org/help/>) and the full backtrace.
 
@@ -118,7 +118,7 @@
       weighted_quantile(1, 1:2, 0.5)
     Condition
       Error in `weighted_quantile()`:
-      ! `x` and `weights` must have the same size.
+      ! `x` (1) and `weights` (2) must have the same size.
 
 # `probabilities` is validated
 
@@ -150,5 +150,5 @@
       weighted_quantile(1, 1, NA)
     Condition
       Error in `weighted_quantile()`:
-      ! `probabilities` can't be missing.
+      ! `probabilities` can't have missing values.
 
diff --git a/tests/testthat/_snaps/num-huber_loss.md b/tests/testthat/_snaps/num-huber_loss.md
index 91108aa1..eaa12251 100644
--- a/tests/testthat/_snaps/num-huber_loss.md
+++ b/tests/testthat/_snaps/num-huber_loss.md
@@ -4,7 +4,7 @@
       huber_loss(ex_dat, truth = "obs", estimate = "pred_na", delta = -1)
     Condition
       Error in `huber_loss()`:
-      ! `delta` must be a positive value.
+      ! `delta` must be a number larger than or equal to 0, not the number -1.
 
 ---
 
@@ -12,5 +12,5 @@
       huber_loss(ex_dat, truth = "obs", estimate = "pred_na", delta = c(1, 2))
     Condition
       Error in `huber_loss()`:
-      ! `delta` must be a single numeric value.
+      ! `delta` must be a number, not a double vector.
 
diff --git a/tests/testthat/_snaps/num-mase.md b/tests/testthat/_snaps/num-mase.md
index 8676bc32..470e39f2 100644
--- a/tests/testthat/_snaps/num-mase.md
+++ b/tests/testthat/_snaps/num-mase.md
@@ -4,7 +4,7 @@
       mase_vec(truth, pred, m = "x")
     Condition
       Error in `mase_vec()`:
-      ! `m` must be a single positive integer value.
+      ! `m` must be a whole number, not the string "x".
 
 ---
 
@@ -12,7 +12,7 @@
       mase_vec(truth, pred, m = -1)
     Condition
       Error in `mase_vec()`:
-      ! `m` must be a single positive integer value.
+      ! `m` must be a whole number larger than or equal to 0, not the number -1.
 
 ---
 
@@ -20,7 +20,7 @@
       mase_vec(truth, pred, m = 1.5)
     Condition
       Error in `mase_vec()`:
-      ! `m` must be a single positive integer value.
+      ! `m` must be a whole number, not the number 1.5.
 
 ---
 
@@ -28,7 +28,7 @@
       mase_vec(truth, pred, mae_train = -1)
     Condition
       Error in `mase_vec()`:
-      ! `mae_train` must be a single positive numeric value.
+      ! `mae_train` must be a number larger than or equal to 0 or `NULL`, not the number -1.
 
 ---
 
@@ -36,5 +36,5 @@
       mase_vec(truth, pred, mae_train = "x")
     Condition
       Error in `mase_vec()`:
-      ! `mae_train` must be a single positive numeric value.
+      ! `mae_train` must be a number or `NULL`, not the string "x".
 
diff --git a/tests/testthat/_snaps/num-pseudo_huber_loss.md b/tests/testthat/_snaps/num-pseudo_huber_loss.md
index 87e8c957..3d0f14f4 100644
--- a/tests/testthat/_snaps/num-pseudo_huber_loss.md
+++ b/tests/testthat/_snaps/num-pseudo_huber_loss.md
@@ -4,7 +4,7 @@
       huber_loss_pseudo(ex_dat, truth = "obs", estimate = "pred_na", delta = -1)
     Condition
       Error in `huber_loss_pseudo()`:
-      ! `delta` must be a positive value.
+      ! `delta` must be a number larger than or equal to 0, not the number -1.
 
 ---
 
@@ -12,5 +12,5 @@
       huber_loss_pseudo(ex_dat, truth = "obs", estimate = "pred_na", delta = c(1, 2))
     Condition
       Error in `huber_loss_pseudo()`:
-      ! `delta` must be a single numeric value.
+      ! `delta` must be a number, not a double vector.
 
diff --git a/tests/testthat/_snaps/prob-average_precision.md b/tests/testthat/_snaps/prob-average_precision.md
index 80a863f5..cabc858b 100644
--- a/tests/testthat/_snaps/prob-average_precision.md
+++ b/tests/testthat/_snaps/prob-average_precision.md
@@ -28,5 +28,5 @@
       average_precision_vec(cp_truth, estimate)
     Condition
       Error in `average_precision_vec()`:
-      ! `truth` should not a `class_pred` object.
+      ! `truth` should not a <class_pred> object.
 
diff --git a/tests/testthat/_snaps/prob-brier_class.md b/tests/testthat/_snaps/prob-brier_class.md
index bdc15b23..98124dd2 100644
--- a/tests/testthat/_snaps/prob-brier_class.md
+++ b/tests/testthat/_snaps/prob-brier_class.md
@@ -4,5 +4,5 @@
       brier_class_vec(cp_truth, estimate)
     Condition
       Error in `brier_class_vec()`:
-      ! `truth` should not a `class_pred` object.
+      ! `truth` should not a <class_pred> object.
 
diff --git a/tests/testthat/_snaps/prob-classification_cost.md b/tests/testthat/_snaps/prob-classification_cost.md
index 4832b4ef..ba5df219 100644
--- a/tests/testthat/_snaps/prob-classification_cost.md
+++ b/tests/testthat/_snaps/prob-classification_cost.md
@@ -12,7 +12,7 @@
       classification_cost(df, obs, A, costs = 1)
     Condition
       Error in `classification_cost()`:
-      ! `costs` must be `NULL` or a data.frame.
+      ! `costs` must be a data frame or `NULL`, not the number 1.
 
 ---
 
@@ -20,7 +20,7 @@
       classification_cost(df, obs, A, costs = data.frame())
     Condition
       Error in `classification_cost()`:
-      ! `costs` must be a data.frame with 3 columns.
+      ! `costs` must be a data.frame with 3 columns, not 0.
 
 ---
 
@@ -28,7 +28,7 @@
       classification_cost(df, obs, A, costs = data.frame(x = 1, y = 2, z = 3))
     Condition
       Error in `classification_cost()`:
-      ! `costs` must have columns: 'truth', 'estimate', and 'cost'.
+      ! `costs` must have columns: "truth", "estimate", and "cost". Not x, y, and z.
 
 # costs$estimate must contain the right levels
 
@@ -52,7 +52,7 @@
       classification_cost(df, obs, A, costs = costs)
     Condition
       Error in `classification_cost()`:
-      ! `costs$truth` must be a character or factor column.
+      ! `costs$truth` must be a character or factor column, not 1 and 2.
 
 ---
 
@@ -60,7 +60,7 @@
       classification_cost(df, obs, A, costs = costs)
     Condition
       Error in `classification_cost()`:
-      ! `costs$estimate` must be a character or factor column.
+      ! `costs$estimate` must be a character or factor column, not 1 and 2.
 
 ---
 
@@ -68,7 +68,7 @@
       classification_cost(df, obs, A, costs = costs)
     Condition
       Error in `classification_cost()`:
-      ! `costs$cost` must be a numeric column.
+      ! `costs$cost` must be a numeric column, not 1 and 2.
 
 # costs$truth and costs$estimate cannot contain duplicate pairs
 
@@ -76,7 +76,7 @@
       classification_cost(df, obs, A, costs = costs)
     Condition
       Error in `classification_cost()`:
-      ! `costs` cannot have duplicate 'truth' / 'estimate' combinations.
+      ! costs cannot have duplicate truth / estimate combinations.
 
 # errors with class_pred input
 
@@ -84,5 +84,5 @@
       classification_cost_vec(cp_truth, estimate)
     Condition
       Error in `classification_cost_vec()`:
-      ! `truth` should not a `class_pred` object.
+      ! `truth` should not a <class_pred> object.
 
diff --git a/tests/testthat/_snaps/prob-gain_capture.md b/tests/testthat/_snaps/prob-gain_capture.md
index 5c0a539f..78dc42ff 100644
--- a/tests/testthat/_snaps/prob-gain_capture.md
+++ b/tests/testthat/_snaps/prob-gain_capture.md
@@ -4,5 +4,5 @@
       gain_capture_vec(cp_truth, estimate)
     Condition
       Error in `gain_capture_vec()`:
-      ! `truth` should not a `class_pred` object.
+      ! `truth` should not a <class_pred> object.
 
diff --git a/tests/testthat/_snaps/prob-lift_curve.md b/tests/testthat/_snaps/prob-lift_curve.md
index 37fea3c1..f03d72f5 100644
--- a/tests/testthat/_snaps/prob-lift_curve.md
+++ b/tests/testthat/_snaps/prob-lift_curve.md
@@ -12,5 +12,5 @@
       lift_curve_vec(cp_truth, estimate)
     Condition
       Error in `gain_curve_vec()`:
-      ! `truth` should not a `class_pred` object.
+      ! `truth` should not a <class_pred> object.
 
diff --git a/tests/testthat/_snaps/prob-mn_log_loss.md b/tests/testthat/_snaps/prob-mn_log_loss.md
index 79c99c85..6f7a7a39 100644
--- a/tests/testthat/_snaps/prob-mn_log_loss.md
+++ b/tests/testthat/_snaps/prob-mn_log_loss.md
@@ -4,5 +4,5 @@
       mn_log_loss_vec(cp_truth, estimate)
     Condition
       Error in `mn_log_loss_vec()`:
-      ! `truth` should not a `class_pred` object.
+      ! `truth` should not a <class_pred> object.
 
diff --git a/tests/testthat/_snaps/prob-pr_auc.md b/tests/testthat/_snaps/prob-pr_auc.md
index ad676c3a..f7f2a014 100644
--- a/tests/testthat/_snaps/prob-pr_auc.md
+++ b/tests/testthat/_snaps/prob-pr_auc.md
@@ -4,5 +4,5 @@
       pr_auc_vec(cp_truth, estimate)
     Condition
       Error in `pr_auc_vec()`:
-      ! `truth` should not a `class_pred` object.
+      ! `truth` should not a <class_pred> object.
 
diff --git a/tests/testthat/_snaps/prob-pr_curve.md b/tests/testthat/_snaps/prob-pr_curve.md
index eb1eb8f7..0338948f 100644
--- a/tests/testthat/_snaps/prob-pr_curve.md
+++ b/tests/testthat/_snaps/prob-pr_curve.md
@@ -12,5 +12,5 @@
       pr_curve_vec(cp_truth, estimate)
     Condition
       Error in `pr_curve_vec()`:
-      ! `truth` should not a `class_pred` object.
+      ! `truth` should not a <class_pred> object.
 
diff --git a/tests/testthat/_snaps/prob-roc_auc.md b/tests/testthat/_snaps/prob-roc_auc.md
index fa41d71d..d291c698 100644
--- a/tests/testthat/_snaps/prob-roc_auc.md
+++ b/tests/testthat/_snaps/prob-roc_auc.md
@@ -109,5 +109,5 @@
       roc_auc_vec(cp_truth, estimate)
     Condition
       Error in `roc_auc_vec()`:
-      ! `truth` should not a `class_pred` object.
+      ! `truth` should not a <class_pred> object.
 
diff --git a/tests/testthat/_snaps/prob-roc_aunp.md b/tests/testthat/_snaps/prob-roc_aunp.md
index ff37629c..04153ead 100644
--- a/tests/testthat/_snaps/prob-roc_aunp.md
+++ b/tests/testthat/_snaps/prob-roc_aunp.md
@@ -31,5 +31,5 @@
       roc_aunp_vec(cp_truth, estimate)
     Condition
       Error in `roc_aunp_vec()`:
-      ! `truth` should not a `class_pred` object.
+      ! `truth` should not a <class_pred> object.
 
diff --git a/tests/testthat/_snaps/prob-roc_aunu.md b/tests/testthat/_snaps/prob-roc_aunu.md
index 7f3f401a..ef22058f 100644
--- a/tests/testthat/_snaps/prob-roc_aunu.md
+++ b/tests/testthat/_snaps/prob-roc_aunu.md
@@ -31,5 +31,5 @@
       roc_aunu_vec(cp_truth, estimate)
     Condition
       Error in `roc_aunu_vec()`:
-      ! `truth` should not a `class_pred` object.
+      ! `truth` should not a <class_pred> object.
 
diff --git a/tests/testthat/_snaps/prob-roc_curve.md b/tests/testthat/_snaps/prob-roc_curve.md
index b3684b10..7e868a4a 100644
--- a/tests/testthat/_snaps/prob-roc_curve.md
+++ b/tests/testthat/_snaps/prob-roc_curve.md
@@ -38,5 +38,5 @@
       roc_curve_vec(cp_truth, estimate)
     Condition
       Error in `roc_curve_vec()`:
-      ! `truth` should not a `class_pred` object.
+      ! `truth` should not a <class_pred> object.
 
diff --git a/tests/testthat/_snaps/probably.md b/tests/testthat/_snaps/probably.md
index dcc7ba86..46e43eeb 100644
--- a/tests/testthat/_snaps/probably.md
+++ b/tests/testthat/_snaps/probably.md
@@ -4,7 +4,7 @@
       accuracy_vec(cp_truth, cp_estimate)
     Condition
       Error in `accuracy_vec()`:
-      ! `truth` should not a `class_pred` object.
+      ! `truth` should not a <class_pred> object.
 
 # `class_pred` errors when passed to `conf_mat()`
 
@@ -12,7 +12,7 @@
       conf_mat(cp_hpc_cv, obs, pred)
     Condition
       Error in `conf_mat_impl()`:
-      ! `truth` should not a `class_pred` object.
+      ! `truth` should not a <class_pred> object.
 
 ---
 
@@ -20,7 +20,7 @@
       conf_mat(dplyr::group_by(cp_hpc_cv, Resample), obs, pred)
     Condition
       Error in `conf_mat_impl()`:
-      ! `truth` should not a `class_pred` object.
+      ! `truth` should not a <class_pred> object.
 
 # `class_pred` errors when passed to `metrics()`
 
@@ -30,5 +30,5 @@
       Error in `metric_set()`:
       ! Failed to compute `accuracy()`.
       Caused by error:
-      ! `truth` should not a `class_pred` object.
+      ! `truth` should not a <class_pred> object.