Slim Frikha
commited on
fix rounding readme
Browse files
README.md
CHANGED
|
@@ -293,7 +293,7 @@ We report in the following table our internal pipeline benchmarks.
|
|
| 293 |
<td>Winogrande (0-shot)</td>
|
| 294 |
<td>-</td>
|
| 295 |
<td>-</td>
|
| 296 |
-
<td>71
|
| 297 |
</tr>
|
| 298 |
<tr>
|
| 299 |
<td>OpenbookQA (0-shot)</td>
|
|
@@ -304,14 +304,14 @@ We report in the following table our internal pipeline benchmarks.
|
|
| 304 |
<tr>
|
| 305 |
<td rowspan="2">Instructions following</td>
|
| 306 |
<td>MT-Bench (avg)</td>
|
| 307 |
-
<td>8.
|
| 308 |
<td><b>8.6</b></td>
|
| 309 |
-
<td>8.
|
| 310 |
</tr>
|
| 311 |
<tr>
|
| 312 |
<td>Alpaca (WC)</td>
|
| 313 |
-
<td>25.
|
| 314 |
-
<td><b>45.
|
| 315 |
<td>24.7</td>
|
| 316 |
</tr>
|
| 317 |
<tr>
|
|
|
|
| 293 |
<td>Winogrande (0-shot)</td>
|
| 294 |
<td>-</td>
|
| 295 |
<td>-</td>
|
| 296 |
+
<td>71</td>
|
| 297 |
</tr>
|
| 298 |
<tr>
|
| 299 |
<td>OpenbookQA (0-shot)</td>
|
|
|
|
| 304 |
<tr>
|
| 305 |
<td rowspan="2">Instructions following</td>
|
| 306 |
<td>MT-Bench (avg)</td>
|
| 307 |
+
<td>8.3</td>
|
| 308 |
<td><b>8.6</b></td>
|
| 309 |
+
<td>8.2</td>
|
| 310 |
</tr>
|
| 311 |
<tr>
|
| 312 |
<td>Alpaca (WC)</td>
|
| 313 |
+
<td>25.8</td>
|
| 314 |
+
<td><b>45.4</b></td>
|
| 315 |
<td>24.7</td>
|
| 316 |
</tr>
|
| 317 |
<tr>
|