Update README.md

README.md

@@ -139,13 +139,36 @@ The training follows casual LLM training practices:
 
 ### Testing Data, Factors & Metrics
 
-#### Metrics
 
-The model is evaluated using standard retrieval metrics:
-- Precision@K (K=2, 5, 10, 20)
-- Recall@K (K=2, 5, 10, 20)
-- Hit Ratio@K (K=2, 5, 10, 20)
-- NDCG@K (K=2, 5, 10, 20)
+### Results
+
+<table>
+<tr>
+<th rowspan="2">Models</th>
+<th colspan="2">Recall@5/10</th>
+<th colspan="2">NDCG@5/10</th>
+</tr>
+<tr>
+<th>Toys and Games</th>
+<th>All Beauty</th>
+<th>Toys and Games</th>
+<th>All Beauty</th>
+</tr>
+<tr>
+<td>PRAG-v1</td>
+<td>0.0452/0.0921</td>
+<td>0.0452/0.0921</td>
+<td>0.0267/0.0414</td>
+<td>0.0267/0.0414</td>
+</tr>
+<tr>
+<td>[TIGER](https://papers.neurips.cc/paper_files/paper/2023/file/20dcab0f14046a5c6b02b61da9f13229-Paper-Conference.pdf)</td>
+<td>0.0531/0.0712</td>
+<td>0.0454/0.0648</td>
+<td>0.0371/0.0432</td>
+<td>0.0321/0.0384</td>
+</tr>
+</table>
 
 ## Technical Specifications