DhiyaEddine
commited on
Update README.md
Browse files
README.md
CHANGED
@@ -110,7 +110,7 @@ We report in the following table our internal pipeline benchmarks:
|
|
110 |
<td>68.7%</td>
|
111 |
<td>-</td>
|
112 |
<td>55.9%</td>
|
113 |
-
<td
|
114 |
</tr>
|
115 |
<tr>
|
116 |
<td>MMLU-PRO (5-shot)</td>
|
@@ -135,7 +135,7 @@ We report in the following table our internal pipeline benchmarks:
|
|
135 |
<td>74.9%</td>
|
136 |
<td>-</td>
|
137 |
<td>19.2%</td>
|
138 |
-
<td
|
139 |
</tr>
|
140 |
<tr>
|
141 |
<td>MATH Lvl-5 (4-shot)</td>
|
@@ -152,7 +152,7 @@ We report in the following table our internal pipeline benchmarks:
|
|
152 |
<td>54.3%</td>
|
153 |
<td>-</td>
|
154 |
<td>46.6%</td>
|
155 |
-
<td
|
156 |
</tr>
|
157 |
<tr>
|
158 |
<td>GPQA (0-shot)</td>
|
@@ -185,7 +185,7 @@ We report in the following table our internal pipeline benchmarks:
|
|
185 |
<td>82.3%</td>
|
186 |
<td>-</td>
|
187 |
<td>78.9%</td>
|
188 |
-
<td
|
189 |
</tr>
|
190 |
<tr>
|
191 |
<td>SciQ (0-shot)</td>
|
@@ -193,7 +193,7 @@ We report in the following table our internal pipeline benchmarks:
|
|
193 |
<td>94.9%</td>
|
194 |
<td>-</td>
|
195 |
<td>80.2%</td>
|
196 |
-
<td
|
197 |
</tr>
|
198 |
<tr>
|
199 |
<td>Winogrande (0-shot)</td>
|
@@ -209,7 +209,7 @@ We report in the following table our internal pipeline benchmarks:
|
|
209 |
<td>34.6%</td>
|
210 |
<td>-</td>
|
211 |
<td>46.2%</td>
|
212 |
-
<td
|
213 |
</tr>
|
214 |
</tbody>
|
215 |
</table>
|
|
|
110 |
<td>68.7%</td>
|
111 |
<td>-</td>
|
112 |
<td>55.9%</td>
|
113 |
+
<td>65.3%</td>
|
114 |
</tr>
|
115 |
<tr>
|
116 |
<td>MMLU-PRO (5-shot)</td>
|
|
|
135 |
<td>74.9%</td>
|
136 |
<td>-</td>
|
137 |
<td>19.2%</td>
|
138 |
+
<td>65.2%</td>
|
139 |
</tr>
|
140 |
<tr>
|
141 |
<td>MATH Lvl-5 (4-shot)</td>
|
|
|
152 |
<td>54.3%</td>
|
153 |
<td>-</td>
|
154 |
<td>46.6%</td>
|
155 |
+
<td>53.7%</td>
|
156 |
</tr>
|
157 |
<tr>
|
158 |
<td>GPQA (0-shot)</td>
|
|
|
185 |
<td>82.3%</td>
|
186 |
<td>-</td>
|
187 |
<td>78.9%</td>
|
188 |
+
<td>80.9%</td>
|
189 |
</tr>
|
190 |
<tr>
|
191 |
<td>SciQ (0-shot)</td>
|
|
|
193 |
<td>94.9%</td>
|
194 |
<td>-</td>
|
195 |
<td>80.2%</td>
|
196 |
+
<td>93.6%</td>
|
197 |
</tr>
|
198 |
<tr>
|
199 |
<td>Winogrande (0-shot)</td>
|
|
|
209 |
<td>34.6%</td>
|
210 |
<td>-</td>
|
211 |
<td>46.2%</td>
|
212 |
+
<td>47.2%</td>
|
213 |
</tr>
|
214 |
</tbody>
|
215 |
</table>
|