diff --git a/spacy/about.py b/spacy/about.py
index 154b7c40c..35e27db7b 100644
--- a/spacy/about.py
+++ b/spacy/about.py
@@ -1,6 +1,6 @@
 # fmt: off
 __title__ = "spacy-nightly"
-__version__ = "3.0.0rc3"
+__version__ = "3.0.0rc4"
 __download_url__ = "https://github.com/explosion/spacy-models/releases/download"
 __compatibility__ = "https://raw.githubusercontent.com/explosion/spacy-models/master/compatibility.json"
 __projects__ = "https://github.com/explosion/projects"
diff --git a/spacy/pipeline/entity_linker.py b/spacy/pipeline/entity_linker.py
index 5ccfe9940..f1ba8637b 100644
--- a/spacy/pipeline/entity_linker.py
+++ b/spacy/pipeline/entity_linker.py
@@ -193,18 +193,16 @@ class EntityLinker(TrainablePipe):
         self,
         examples: Iterable[Example],
         *,
-        set_annotations: bool = False,
         drop: float = 0.0,
         sgd: Optional[Optimizer] = None,
         losses: Optional[Dict[str, float]] = None,
     ) -> Dict[str, float]:
         """Learn from a batch of documents and gold-standard information,
-        updating the pipe's model. Delegates to predict and get_loss.
+        updating the pipe's model. Delegates to predict, get_loss and
+        set_annotations.
 
         examples (Iterable[Example]): A batch of Example objects.
         drop (float): The dropout rate.
-        set_annotations (bool): Whether or not to update the Example objects
-            with the predictions.
         sgd (thinc.api.Optimizer): The optimizer.
         losses (Dict[str, float]): Optional record of the loss during training.
             Updated using the component name as the key.
@@ -220,11 +218,13 @@ class EntityLinker(TrainablePipe):
             return losses
         validate_examples(examples, "EntityLinker.update")
         sentence_docs = []
-        docs = [eg.predicted for eg in examples]
-        if set_annotations:
-            # This seems simpler than other ways to get that exact output -- but
-            # it does run the model twice :(
-            predictions = self.model.predict(docs)
+        docs = []
+        for eg in examples:
+            eg.predicted.ents = eg.reference.ents
+            docs.append(eg.predicted)
+        # This seems simpler than other ways to get that exact output -- but
+        # it does run the model twice :(
+        predictions = self.predict(docs)
         for eg in examples:
             sentences = [s for s in eg.reference.sents]
             kb_ids = eg.get_aligned("ENT_KB_ID", as_string=True)
@@ -260,8 +260,7 @@ class EntityLinker(TrainablePipe):
         if sgd is not None:
             self.finish_update(sgd)
         losses[self.name] += loss
-        if set_annotations:
-            self.set_annotations(docs, predictions)
+        self.set_annotations(docs, predictions)
         return losses
 
     def get_loss(self, examples: Iterable[Example], sentence_encodings):
diff --git a/spacy/pipeline/multitask.pyx b/spacy/pipeline/multitask.pyx
index 9c7bb5914..cfb492612 100644
--- a/spacy/pipeline/multitask.pyx
+++ b/spacy/pipeline/multitask.pyx
@@ -199,7 +199,7 @@ class ClozeMultitask(TrainablePipe):
         loss = self.distance.get_loss(prediction, target)
         return loss, gradient
 
-    def update(self, examples, *, drop=0., set_annotations=False, sgd=None, losses=None):
+    def update(self, examples, *, drop=0., sgd=None, losses=None):
         pass
 
     def rehearse(self, examples, drop=0., sgd=None, losses=None):
diff --git a/spacy/pipeline/tagger.pyx b/spacy/pipeline/tagger.pyx
index d3f8c339c..85592aba5 100644
--- a/spacy/pipeline/tagger.pyx
+++ b/spacy/pipeline/tagger.pyx
@@ -173,14 +173,13 @@ class Tagger(TrainablePipe):
                 if doc.c[j].tag == 0:
                     doc.c[j].tag = self.vocab.strings[self.labels[tag_id]]
 
-    def update(self, examples, *, drop=0., sgd=None, losses=None, set_annotations=False):
+    def update(self, examples, *, drop=0., sgd=None, losses=None):
         """Learn from a batch of documents and gold-standard information,
-        updating the pipe's model. Delegates to predict and get_loss.
+        updating the pipe's model. Delegates to predict, get_loss and
+        set_annotations.
 
         examples (Iterable[Example]): A batch of Example objects.
         drop (float): The dropout rate.
-        set_annotations (bool): Whether or not to update the Example objects
-            with the predictions.
         sgd (thinc.api.Optimizer): The optimizer.
         losses (Dict[str, float]): Optional record of the loss during training.
             Updated using the component name as the key.
@@ -206,9 +205,8 @@ class Tagger(TrainablePipe):
             self.finish_update(sgd)
 
         losses[self.name] += loss
-        if set_annotations:
-            docs = [eg.predicted for eg in examples]
-            self.set_annotations(docs, self._scores2guesses(tag_scores))
+        docs = [eg.predicted for eg in examples]
+        self.set_annotations(docs, self._scores2guesses(tag_scores))
         return losses
 
     def rehearse(self, examples, *, drop=0., sgd=None, losses=None):
diff --git a/spacy/pipeline/textcat.py b/spacy/pipeline/textcat.py
index 79b6e8d38..6d8c7b101 100644
--- a/spacy/pipeline/textcat.py
+++ b/spacy/pipeline/textcat.py
@@ -195,17 +195,15 @@ class TextCategorizer(TrainablePipe):
         examples: Iterable[Example],
         *,
         drop: float = 0.0,
-        set_annotations: bool = False,
         sgd: Optional[Optimizer] = None,
         losses: Optional[Dict[str, float]] = None,
     ) -> Dict[str, float]:
         """Learn from a batch of documents and gold-standard information,
-        updating the pipe's model. Delegates to predict and get_loss.
+        updating the pipe's model. Delegates to predict, get_loss and
+        set_annotations.
 
         examples (Iterable[Example]): A batch of Example objects.
         drop (float): The dropout rate.
-        set_annotations (bool): Whether or not to update the Example objects
-            with the predictions.
         sgd (thinc.api.Optimizer): The optimizer.
         losses (Dict[str, float]): Optional record of the loss during training.
             Updated using the component name as the key.
@@ -228,9 +226,8 @@ class TextCategorizer(TrainablePipe):
         if sgd is not None:
             self.finish_update(sgd)
         losses[self.name] += loss
-        if set_annotations:
-            docs = [eg.predicted for eg in examples]
-            self.set_annotations(docs, scores=scores)
+        docs = [eg.predicted for eg in examples]
+        self.set_annotations(docs, scores=scores)
         return losses
 
     def rehearse(
diff --git a/spacy/pipeline/tok2vec.py b/spacy/pipeline/tok2vec.py
index 1220611b8..a95fc8927 100644
--- a/spacy/pipeline/tok2vec.py
+++ b/spacy/pipeline/tok2vec.py
@@ -163,15 +163,12 @@ class Tok2Vec(TrainablePipe):
         drop: float = 0.0,
         sgd: Optional[Optimizer] = None,
         losses: Optional[Dict[str, float]] = None,
-        set_annotations: bool = False,
     ):
         """Learn from a batch of documents and gold-standard information,
         updating the pipe's model.
 
         examples (Iterable[Example]): A batch of Example objects.
         drop (float): The dropout rate.
-        set_annotations (bool): Whether or not to update the Example objects
-            with the predictions.
         sgd (thinc.api.Optimizer): The optimizer.
         losses (Dict[str, float]): Optional record of the loss during training.
             Updated using the component name as the key.
@@ -210,8 +207,7 @@ class Tok2Vec(TrainablePipe):
             listener.receive(batch_id, tokvecs, accumulate_gradient)
         if self.listeners:
             self.listeners[-1].receive(batch_id, tokvecs, backprop)
-        if set_annotations:
-            self.set_annotations(docs, tokvecs)
+        self.set_annotations(docs, tokvecs)
         return losses
 
     def get_loss(self, examples, scores) -> None:
diff --git a/spacy/pipeline/trainable_pipe.pyx b/spacy/pipeline/trainable_pipe.pyx
index 6cd73d256..1abd6b43e 100644
--- a/spacy/pipeline/trainable_pipe.pyx
+++ b/spacy/pipeline/trainable_pipe.pyx
@@ -91,16 +91,14 @@ cdef class TrainablePipe(Pipe):
     def update(self,
                examples: Iterable["Example"],
                *, drop: float=0.0,
-               set_annotations: bool=False,
                sgd: Optimizer=None,
                losses: Optional[Dict[str, float]]=None) -> Dict[str, float]:
         """Learn from a batch of documents and gold-standard information,
-        updating the pipe's model. Delegates to predict and get_loss.
+        updating the pipe's model. Delegates to predict, get_loss and
+        set_annotations.
 
         examples (Iterable[Example]): A batch of Example objects.
         drop (float): The dropout rate.
-        set_annotations (bool): Whether or not to update the Example objects
-            with the predictions.
         sgd (thinc.api.Optimizer): The optimizer.
         losses (Dict[str, float]): Optional record of the loss during training.
             Updated using the component name as the key.
@@ -124,9 +122,8 @@ cdef class TrainablePipe(Pipe):
         if sgd not in (None, False):
             self.finish_update(sgd)
         losses[self.name] += loss
-        if set_annotations:
-            docs = [eg.predicted for eg in examples]
-            self.set_annotations(docs, scores=scores)
+        docs = [eg.predicted for eg in examples]
+        self.set_annotations(docs, scores=scores)
         return losses
 
     def rehearse(self,
diff --git a/spacy/pipeline/transition_parser.pyx b/spacy/pipeline/transition_parser.pyx
index 8aeacbafb..8cb4ea15d 100644
--- a/spacy/pipeline/transition_parser.pyx
+++ b/spacy/pipeline/transition_parser.pyx
@@ -308,7 +308,7 @@ cdef class Parser(TrainablePipe):
                 action.do(states[i], action.label)
         free(is_valid)
 
-    def update(self, examples, *, drop=0., set_annotations=False, sgd=None, losses=None):
+    def update(self, examples, *, drop=0., sgd=None, losses=None):
         cdef StateClass state
         if losses is None:
             losses = {}
@@ -328,7 +328,6 @@ cdef class Parser(TrainablePipe):
             return self.update_beam(
                 examples,
                 beam_width=self.cfg["beam_width"],
-                set_annotations=set_annotations,
                 sgd=sgd,
                 losses=losses,
                 beam_density=self.cfg["beam_density"]
@@ -370,9 +369,8 @@ cdef class Parser(TrainablePipe):
         backprop_tok2vec(golds)
         if sgd not in (None, False):
             self.finish_update(sgd)
-        if set_annotations:
-            docs = [eg.predicted for eg in examples]
-            self.set_annotations(docs, all_states)
+        docs = [eg.predicted for eg in examples]
+        self.set_annotations(docs, all_states)
         # Ugh, this is annoying. If we're working on GPU, we want to free the
         # memory ASAP. It seems that Python doesn't necessarily get around to
         # removing these in time if we don't explicitly delete? It's confusing.
@@ -432,7 +430,7 @@ cdef class Parser(TrainablePipe):
         return losses
 
     def update_beam(self, examples, *, beam_width,
-            drop=0., sgd=None, losses=None, set_annotations=False, beam_density=0.0):
+            drop=0., sgd=None, losses=None, beam_density=0.0):
         states, golds, _ = self.moves.init_gold_batch(examples)
         if not states:
             return losses
diff --git a/spacy/tests/pipeline/test_entity_linker.py b/spacy/tests/pipeline/test_entity_linker.py
index 8ba2d0d3e..743800536 100644
--- a/spacy/tests/pipeline/test_entity_linker.py
+++ b/spacy/tests/pipeline/test_entity_linker.py
@@ -425,6 +425,7 @@ GOLD_entities = ["Q2146908", "Q7381115", "Q7381115", "Q2146908"]
 def test_overfitting_IO():
     # Simple test to try and quickly overfit the NEL component - ensuring the ML models work correctly
     nlp = English()
+    nlp.add_pipe("sentencizer", first=True)
     vector_length = 3
     assert "Q2146908" not in nlp.vocab.strings
 
@@ -464,9 +465,6 @@ def test_overfitting_IO():
         nlp.update(train_examples, sgd=optimizer, losses=losses)
     assert losses["entity_linker"] < 0.001
 
-    # adding additional components that are required for the entity_linker
-    nlp.add_pipe("sentencizer", first=True)
-
     # Add a custom component to recognize "Russ Cochran" as an entity for the example training data
     patterns = [
         {"label": "PERSON", "pattern": [{"LOWER": "russ"}, {"LOWER": "cochran"}]}
diff --git a/website/docs/api/dependencyparser.md b/website/docs/api/dependencyparser.md
index fe8f7d8d5..8974d9ea7 100644
--- a/website/docs/api/dependencyparser.md
+++ b/website/docs/api/dependencyparser.md
@@ -220,8 +220,9 @@ Modify a batch of [`Doc`](/api/doc) objects, using pre-computed scores.
 ## DependencyParser.update {#update tag="method"}
 
 Learn from a batch of [`Example`](/api/example) objects, updating the pipe's
-model. Delegates to [`predict`](/api/dependencyparser#predict) and
-[`get_loss`](/api/dependencyparser#get_loss).
+model. Delegates to [`predict`](/api/dependencyparser#predict), 
+[`get_loss`](/api/dependencyparser#get_loss) and 
+[`set_annotations`](/api/dependencyparser#set_annotations).
 
 > #### Example
 >
@@ -236,7 +237,6 @@ model. Delegates to [`predict`](/api/dependencyparser#predict) and
 | `examples`        | A batch of [`Example`](/api/example) objects to learn from. ~~Iterable[Example]~~                                                  |
 | _keyword-only_    |                                                                                                                                    |
 | `drop`            | The dropout rate. ~~float~~                                                                                                        |
-| `set_annotations` | Whether or not to update the `Example` objects with the predictions, delegating to [`set_annotations`](#set_annotations). ~~bool~~ |
 | `sgd`             | An optimizer. Will be created via [`create_optimizer`](#create_optimizer) if not set. ~~Optional[Optimizer]~~                      |
 | `losses`          | Optional record of the loss during training. Updated using the component name as the key. ~~Optional[Dict[str, float]]~~           |
 | **RETURNS**       | The updated `losses` dictionary. ~~Dict[str, float]~~                                                                              |
diff --git a/website/docs/api/entitylinker.md b/website/docs/api/entitylinker.md
index 683927b1c..b90c52710 100644
--- a/website/docs/api/entitylinker.md
+++ b/website/docs/api/entitylinker.md
@@ -237,7 +237,8 @@ entities.
 
 Learn from a batch of [`Example`](/api/example) objects, updating both the
 pipe's entity linking model and context encoder. Delegates to
-[`predict`](/api/entitylinker#predict).
+[`predict`](/api/entitylinker#predict) and 
+[`set_annotations`](/api/entitylinker#set_annotations).
 
 > #### Example
 >
@@ -252,7 +253,6 @@ pipe's entity linking model and context encoder. Delegates to
 | `examples`        | A batch of [`Example`](/api/example) objects to learn from. ~~Iterable[Example]~~                                                  |
 | _keyword-only_    |                                                                                                                                    |
 | `drop`            | The dropout rate. ~~float~~                                                                                                        |
-| `set_annotations` | Whether or not to update the `Example` objects with the predictions, delegating to [`set_annotations`](#set_annotations). ~~bool~~ |
 | `sgd`             | An optimizer. Will be created via [`create_optimizer`](#create_optimizer) if not set. ~~Optional[Optimizer]~~                      |
 | `losses`          | Optional record of the loss during training. Updated using the component name as the key. ~~Optional[Dict[str, float]]~~           |
 | **RETURNS**       | The updated `losses` dictionary. ~~Dict[str, float]~~                                                                              |
diff --git a/website/docs/api/entityrecognizer.md b/website/docs/api/entityrecognizer.md
index 6ac0d163f..dd969d14b 100644
--- a/website/docs/api/entityrecognizer.md
+++ b/website/docs/api/entityrecognizer.md
@@ -209,8 +209,9 @@ Modify a batch of [`Doc`](/api/doc) objects, using pre-computed scores.
 ## EntityRecognizer.update {#update tag="method"}
 
 Learn from a batch of [`Example`](/api/example) objects, updating the pipe's
-model. Delegates to [`predict`](/api/entityrecognizer#predict) and
-[`get_loss`](/api/entityrecognizer#get_loss).
+model. Delegates to [`predict`](/api/entityrecognizer#predict), 
+[`get_loss`](/api/entityrecognizer#get_loss) and 
+[`set_annotations`](/api/entityrecognizer#set_annotations).
 
 > #### Example
 >
@@ -225,7 +226,6 @@ model. Delegates to [`predict`](/api/entityrecognizer#predict) and
 | `examples`        | A batch of [`Example`](/api/example) objects to learn from. ~~Iterable[Example]~~                                                  |
 | _keyword-only_    |                                                                                                                                    |
 | `drop`            | The dropout rate. ~~float~~                                                                                                        |
-| `set_annotations` | Whether or not to update the `Example` objects with the predictions, delegating to [`set_annotations`](#set_annotations). ~~bool~~ |
 | `sgd`             | An optimizer. Will be created via [`create_optimizer`](#create_optimizer) if not set. ~~Optional[Optimizer]~~                      |
 | `losses`          | Optional record of the loss during training. Updated using the component name as the key. ~~Optional[Dict[str, float]]~~           |
 | **RETURNS**       | The updated `losses` dictionary. ~~Dict[str, float]~~                                                                              |
diff --git a/website/docs/api/morphologizer.md b/website/docs/api/morphologizer.md
index fed86ff5d..9cda478c8 100644
--- a/website/docs/api/morphologizer.md
+++ b/website/docs/api/morphologizer.md
@@ -189,8 +189,9 @@ Modify a batch of [`Doc`](/api/doc) objects, using pre-computed scores.
 
 Learn from a batch of [`Example`](/api/example) objects containing the
 predictions and gold-standard annotations, and update the component's model.
-Delegates to [`predict`](/api/morphologizer#predict) and
-[`get_loss`](/api/morphologizer#get_loss).
+Delegates to [`predict`](/api/morphologizer#predict), 
+[`get_loss`](/api/morphologizer#get_loss) and 
+[`set_annotations`](/api/morphologizer#set_annotations).
 
 > #### Example
 >
@@ -205,7 +206,6 @@ Delegates to [`predict`](/api/morphologizer#predict) and
 | `examples`        | A batch of [`Example`](/api/example) objects to learn from. ~~Iterable[Example]~~                                                  |
 | _keyword-only_    |                                                                                                                                    |
 | `drop`            | The dropout rate. ~~float~~                                                                                                        |
-| `set_annotations` | Whether or not to update the `Example` objects with the predictions, delegating to [`set_annotations`](#set_annotations). ~~bool~~ |
 | `sgd`             | An optimizer. Will be created via [`create_optimizer`](#create_optimizer) if not set. ~~Optional[Optimizer]~~                      |
 | `losses`          | Optional record of the loss during training. Updated using the component name as the key. ~~Optional[Dict[str, float]]~~           |
 | **RETURNS**       | The updated `losses` dictionary. ~~Dict[str, float]~~                                                                              |
diff --git a/website/docs/api/multilabel_textcategorizer.md b/website/docs/api/multilabel_textcategorizer.md
index 8400c68f7..d74f7ad9d 100644
--- a/website/docs/api/multilabel_textcategorizer.md
+++ b/website/docs/api/multilabel_textcategorizer.md
@@ -199,8 +199,9 @@ Modify a batch of [`Doc`](/api/doc) objects using pre-computed scores.
 
 Learn from a batch of [`Example`](/api/example) objects containing the
 predictions and gold-standard annotations, and update the component's model.
-Delegates to [`predict`](/api/multilabel_textcategorizer#predict) and
-[`get_loss`](/api/multilabel_textcategorizer#get_loss).
+Delegates to [`predict`](/api/multilabel_textcategorizer#predict),
+[`get_loss`](/api/multilabel_textcategorizer#get_loss) and 
+[`set_annotations`](/api/multilabel_textcategorizer#set_annotations).
 
 > #### Example
 >
@@ -215,7 +216,6 @@ Delegates to [`predict`](/api/multilabel_textcategorizer#predict) and
 | `examples`        | A batch of [`Example`](/api/example) objects to learn from. ~~Iterable[Example]~~                                                  |
 | _keyword-only_    |                                                                                                                                    |
 | `drop`            | The dropout rate. ~~float~~                                                                                                        |
-| `set_annotations` | Whether or not to update the `Example` objects with the predictions, delegating to [`set_annotations`](#set_annotations). ~~bool~~ |
 | `sgd`             | An optimizer. Will be created via [`create_optimizer`](#create_optimizer) if not set. ~~Optional[Optimizer]~~                      |
 | `losses`          | Optional record of the loss during training. Updated using the component name as the key. ~~Optional[Dict[str, float]]~~           |
 | **RETURNS**       | The updated `losses` dictionary. ~~Dict[str, float]~~                                                                              |
diff --git a/website/docs/api/pipe.md b/website/docs/api/pipe.md
index 1f7fab8aa..027d1e1c0 100644
--- a/website/docs/api/pipe.md
+++ b/website/docs/api/pipe.md
@@ -195,7 +195,6 @@ predictions and gold-standard annotations, and update the component's model.
 | `examples`        | A batch of [`Example`](/api/example) objects to learn from. ~~Iterable[Example]~~                                                  |
 | _keyword-only_    |                                                                                                                                    |
 | `drop`            | The dropout rate. ~~float~~                                                                                                        |
-| `set_annotations` | Whether or not to update the `Example` objects with the predictions, delegating to [`set_annotations`](#set_annotations). ~~bool~~ |
 | `sgd`             | An optimizer. Will be created via [`create_optimizer`](#create_optimizer) if not set. ~~Optional[Optimizer]~~                      |
 | `losses`          | Optional record of the loss during training. Updated using the component name as the key. ~~Optional[Dict[str, float]]~~           |
 | **RETURNS**       | The updated `losses` dictionary. ~~Dict[str, float]~~                                                                              |
diff --git a/website/docs/api/sentencerecognizer.md b/website/docs/api/sentencerecognizer.md
index fced37fd3..23c8e87d9 100644
--- a/website/docs/api/sentencerecognizer.md
+++ b/website/docs/api/sentencerecognizer.md
@@ -176,8 +176,9 @@ Modify a batch of [`Doc`](/api/doc) objects, using pre-computed scores.
 
 Learn from a batch of [`Example`](/api/example) objects containing the
 predictions and gold-standard annotations, and update the component's model.
-Delegates to [`predict`](/api/sentencerecognizer#predict) and
-[`get_loss`](/api/sentencerecognizer#get_loss).
+Delegates to [`predict`](/api/sentencerecognizer#predict), 
+[`get_loss`](/api/sentencerecognizer#get_loss) and 
+[`set_annotations`](/api/sentencerecognizer#set_annotations).
 
 > #### Example
 >
@@ -192,7 +193,6 @@ Delegates to [`predict`](/api/sentencerecognizer#predict) and
 | `examples`        | A batch of [`Example`](/api/example) objects to learn from. ~~Iterable[Example]~~                                                  |
 | _keyword-only_    |                                                                                                                                    |
 | `drop`            | The dropout rate. ~~float~~                                                                                                        |
-| `set_annotations` | Whether or not to update the `Example` objects with the predictions, delegating to [`set_annotations`](#set_annotations). ~~bool~~ |
 | `sgd`             | An optimizer. Will be created via [`create_optimizer`](#create_optimizer) if not set. ~~Optional[Optimizer]~~                      |
 | `losses`          | Optional record of the loss during training. Updated using the component name as the key. ~~Optional[Dict[str, float]]~~           |
 | **RETURNS**       | The updated `losses` dictionary. ~~Dict[str, float]~~                                                                              |
diff --git a/website/docs/api/tagger.md b/website/docs/api/tagger.md
index f337b51fd..8e6132d40 100644
--- a/website/docs/api/tagger.md
+++ b/website/docs/api/tagger.md
@@ -187,8 +187,9 @@ Modify a batch of [`Doc`](/api/doc) objects, using pre-computed scores.
 
 Learn from a batch of [`Example`](/api/example) objects containing the
 predictions and gold-standard annotations, and update the component's model.
-Delegates to [`predict`](/api/tagger#predict) and
-[`get_loss`](/api/tagger#get_loss).
+Delegates to [`predict`](/api/tagger#predict), 
+[`get_loss`](/api/tagger#get_loss) and 
+[`set_annotations`](/api/tagger#set_annotations).
 
 > #### Example
 >
@@ -203,7 +204,6 @@ Delegates to [`predict`](/api/tagger#predict) and
 | `examples`        | A batch of [`Example`](/api/example) objects to learn from. ~~Iterable[Example]~~                                                  |
 | _keyword-only_    |                                                                                                                                    |
 | `drop`            | The dropout rate. ~~float~~                                                                                                        |
-| `set_annotations` | Whether or not to update the `Example` objects with the predictions, delegating to [`set_annotations`](#set_annotations). ~~bool~~ |
 | `sgd`             | An optimizer. Will be created via [`create_optimizer`](#create_optimizer) if not set. ~~Optional[Optimizer]~~                      |
 | `losses`          | Optional record of the loss during training. Updated using the component name as the key. ~~Optional[Dict[str, float]]~~           |
 | **RETURNS**       | The updated `losses` dictionary. ~~Dict[str, float]~~                                                                              |
diff --git a/website/docs/api/textcategorizer.md b/website/docs/api/textcategorizer.md
index 016b9c310..16049c327 100644
--- a/website/docs/api/textcategorizer.md
+++ b/website/docs/api/textcategorizer.md
@@ -201,8 +201,9 @@ Modify a batch of [`Doc`](/api/doc) objects using pre-computed scores.
 
 Learn from a batch of [`Example`](/api/example) objects containing the
 predictions and gold-standard annotations, and update the component's model.
-Delegates to [`predict`](/api/textcategorizer#predict) and
-[`get_loss`](/api/textcategorizer#get_loss).
+Delegates to [`predict`](/api/textcategorizer#predict),
+[`get_loss`](/api/textcategorizer#get_loss) and 
+[`set_annotations`](/api/textcategorizer#set_annotations).
 
 > #### Example
 >
@@ -217,7 +218,6 @@ Delegates to [`predict`](/api/textcategorizer#predict) and
 | `examples`        | A batch of [`Example`](/api/example) objects to learn from. ~~Iterable[Example]~~                                                  |
 | _keyword-only_    |                                                                                                                                    |
 | `drop`            | The dropout rate. ~~float~~                                                                                                        |
-| `set_annotations` | Whether or not to update the `Example` objects with the predictions, delegating to [`set_annotations`](#set_annotations). ~~bool~~ |
 | `sgd`             | An optimizer. Will be created via [`create_optimizer`](#create_optimizer) if not set. ~~Optional[Optimizer]~~                      |
 | `losses`          | Optional record of the loss during training. Updated using the component name as the key. ~~Optional[Dict[str, float]]~~           |
 | **RETURNS**       | The updated `losses` dictionary. ~~Dict[str, float]~~                                                                              |
diff --git a/website/docs/api/tok2vec.md b/website/docs/api/tok2vec.md
index 051164ff5..409c7f25b 100644
--- a/website/docs/api/tok2vec.md
+++ b/website/docs/api/tok2vec.md
@@ -186,7 +186,8 @@ Modify a batch of [`Doc`](/api/doc) objects, using pre-computed scores.
 
 Learn from a batch of [`Example`](/api/example) objects containing the
 predictions and gold-standard annotations, and update the component's model.
-Delegates to [`predict`](/api/tok2vec#predict).
+Delegates to [`predict`](/api/tok2vec#predict) and 
+[`set_annotations`](/api/tok2vec#set_annotations).
 
 > #### Example
 >
@@ -201,7 +202,6 @@ Delegates to [`predict`](/api/tok2vec#predict).
 | `examples`        | A batch of [`Example`](/api/example) objects to learn from. ~~Iterable[Example]~~                                                  |
 | _keyword-only_    |                                                                                                                                    |
 | `drop`            | The dropout rate. ~~float~~                                                                                                        |
-| `set_annotations` | Whether or not to update the `Example` objects with the predictions, delegating to [`set_annotations`](#set_annotations). ~~bool~~ |
 | `sgd`             | An optimizer. Will be created via [`create_optimizer`](#create_optimizer) if not set. ~~Optional[Optimizer]~~                      |
 | `losses`          | Optional record of the loss during training. Updated using the component name as the key. ~~Optional[Dict[str, float]]~~           |
 | **RETURNS**       | The updated `losses` dictionary. ~~Dict[str, float]~~                                                                              |
diff --git a/website/docs/api/transformer.md b/website/docs/api/transformer.md
index e31c8ad2c..a38b95a8b 100644
--- a/website/docs/api/transformer.md
+++ b/website/docs/api/transformer.md
@@ -249,7 +249,6 @@ and call the optimizer, while the others simply increment the gradients.
 | `examples`        | A batch of [`Example`](/api/example) objects. Only the [`Example.predicted`](/api/example#predicted) `Doc` object is used, the reference `Doc` is ignored. ~~Iterable[Example]~~ |
 | _keyword-only_    |                                                                                                                                                                                  |
 | `drop`            | The dropout rate. ~~float~~                                                                                                                                                      |
-| `set_annotations` | Whether or not to update the `Example` objects with the predictions, delegating to [`set_annotations`](#set_annotations). ~~bool~~                                               |
 | `sgd`             | An optimizer. Will be created via [`create_optimizer`](#create_optimizer) if not set. ~~Optional[Optimizer]~~                                                                    |
 | `losses`          | Optional record of the loss during training. Updated using the component name as the key. ~~Optional[Dict[str, float]]~~                                                         |
 | **RETURNS**       | The updated `losses` dictionary. ~~Dict[str, float]~~                                                                                                                            |
diff --git a/website/docs/usage/layers-architectures.md b/website/docs/usage/layers-architectures.md
index 33f647413..13e474bfe 100644
--- a/website/docs/usage/layers-architectures.md
+++ b/website/docs/usage/layers-architectures.md
@@ -796,7 +796,7 @@ class RelationExtractor(TrainablePipe):
         self.vocab = vocab
         self.name = name
 
-    def update(self, examples, drop=0.0, set_annotations=False, sgd=None, losses=None):
+    def update(self, examples, drop=0.0, sgd=None, losses=None):
         """Learn from a batch of Example objects."""
         ...
 
@@ -901,7 +901,6 @@ def update(
     examples: Iterable[Example],
     *,
     drop: float = 0.0,
-    set_annotations: bool = False,
     sgd: Optional[Optimizer] = None,
     losses: Optional[Dict[str, float]] = None,
 ) -> Dict[str, float]: