Datasets:

Omartificial-Intelligence-Space
/

Arabic-With-Ranked-Hard-Negatives

@@ -95,7 +95,7 @@ Languages: Arabic (Modern Standard Arabic).
 ## Dataset Analysis and Insights
-1. Average Similarity Across Negatives:
 ![Gate-sim-results](https://i.ibb.co/7SKdT2F/Gate-sim-results.png)
@@ -103,11 +103,11 @@ Languages: Arabic (Modern Standard Arabic).
 ![Gate-sim-results-dis](https://i.ibb.co/gTQD4GH/Gate-sim-result-dis.png)
-2. Similarity Distributions:
 🔸The similarity scores for each negative passage are distributed differently. Below are the histograms for the similarity distributions of the top 30 negatives, emphasizing the scores for negative1 to negative4.
-3. Insights
 The top-ranked negatives (negative1 and negative2) are significantly closer in similarity to the positive passage, making them challenging and ideal for training advanced retrieval models.
 The similarity drops slightly for negative3 and negative4, but they remain "hard negatives," offering diverse yet challenging non-relevant passages for contrastive learning.

 ## Dataset Analysis and Insights
+### 1. Average Similarity Across Negatives:
 ![Gate-sim-results](https://i.ibb.co/7SKdT2F/Gate-sim-results.png)
 ![Gate-sim-results-dis](https://i.ibb.co/gTQD4GH/Gate-sim-result-dis.png)
+### 2. Similarity Distributions:
 🔸The similarity scores for each negative passage are distributed differently. Below are the histograms for the similarity distributions of the top 30 negatives, emphasizing the scores for negative1 to negative4.
+### 3. Insights
 The top-ranked negatives (negative1 and negative2) are significantly closer in similarity to the positive passage, making them challenging and ideal for training advanced retrieval models.
 The similarity drops slightly for negative3 and negative4, but they remain "hard negatives," offering diverse yet challenging non-relevant passages for contrastive learning.