diff --git a/docs/release-notes.md b/docs/release-notes.md
index a2bf798cb..ad5c36d5e 100644
--- a/docs/release-notes.md
+++ b/docs/release-notes.md
@@ -261,12 +261,12 @@ Get started by selecting an AI workflow you want to evaluate that includes one o
 
 1. Add an **Evaluation** node with the **On new Evaluation event** operation. This node will act as an additional trigger you’ll run only when testing. Configure it to read your dataset from Google Sheets, with each row representing a test input.<br>
 
-    > 💡  Better datasets mean better evaluations. Craft your dataset from a variety of test cases, including edge cases and typical inputs, to get meaningful feedback on how your AI performs. Learn more and access sample datasets [here](/advanced-ai/evaluations/light-evaluations.md/#1-create-a-dataset.md).
+    > 💡  Better datasets mean better evaluations. Craft your dataset from a variety of test cases, including edge cases and typical inputs, to get meaningful feedback on how your AI performs. Learn more and access sample datasets [here](/advanced-ai/evaluations/light-evaluations.md/#1-create-a-dataset).
 
 2. Add a second **Evaluation** node using the **Set Outputs** operation after the part of the workflow you're testing—typically after an LLM or Agent node. This captures the response and writes it back to your dataset in Google Sheets.
 3. To evaluate output quality, add a third **Evaluation** node with the **Set Metrics** operation at a point after you’ve generated the outputs. You can develop workflow logic, custom calculations, or add an LLM-as-Judge to score the outputs. Map these metrics to your dataset in the node’s parameters. <br>
 
-    > 💡 Well-defined metrics = smarter decisions. Scoring your outputs based on similarity, correctness, or categorization can help you track whether changes are actually improving performance. Learn more and get links to example templates [here](/advanced-ai/evaluations/metric-based-evaluations.md/#2-calculate-metrics.md). 
+    > 💡 Well-defined metrics = smarter decisions. Scoring your outputs based on similarity, correctness, or categorization can help you track whether changes are actually improving performance. Learn more and get links to example templates [here](/advanced-ai/evaluations/metric-based-evaluations.md/#2-calculate-metrics). 
     
 <br>