docs: fix multiple typos in gemma4_unified_assistant.md

docs/source/en/model_doc/gemma4_unified_assistant.md

@@ -23,17 +23,17 @@ limitations under the License.
 
 ## Overview
 
-Gemma 4 Unified Assistant is a small, text-only model that enables speculative decoding with for Gemma 4 Unified models using the
+Gemma 4 Unified Assistant is a small, text-only model that enables speculative decoding for Gemma 4 Unified models using the
 Multi-Token Prediction (MTP) method and associated candidate generator. Pre-trained models are provided for the IT
 variants of the Gemma 4 12B model.
 
 For more information, please see [Gemma4 Assistant](./gemma4_assistant.md). Architecturally and conceptually, they
 share the same concept and differences to their base model:
 
 *   **The entire model uses KV sharing**. This technique, originally introduced with [Gemma 3n](./gemma3n), allows the
-    model to resuse the KV cache populated by the target model the assistant supports, allowing the assistant to skip
-    the pre-fille phase entirely, and considerably reducing attention compute during the forward pass.
-*   **The `position_ids` value are constant**. Since the KV cache is shared and the assistant does not have a mean of
+    model to reuse the KV cache populated by the target model the assistant supports, allowing the assistant to skip
+    the pre-fill phase entirely, and considerably reducing attention compute during the forward pass.
+*   **The `position_ids` value are constant**. Since the KV cache is shared and the assistant does not have a means of
     updating the cache, the assistant predicts all tokens from the same position ID.
 *   **Inputs are the concatenation of embeddings and hidden states**. To adapt for the static KV cache and
     `position_ids`, the model takes its inputs as the concatenation of the `embedding` and `hidden_states` for the last
@@ -43,7 +43,7 @@ share the same concept and differences to their base model:
     token will be the last token generated by the assistant (within a drafting round) or the last token accepted by the
     target model (between drafting rounds).
 *   **Cross-attention is used to make the most of the target model's context**. Cross-attention allows the query states
-    geneated by the assistant to attend to the shared KV cache values from the target model, allowing the assistant to
+    generated by the assistant to attend to the shared KV cache values from the target model, allowing the assistant to
     accurately predict more drafted tokens per drafting round.
 
 ## Usage examples