Update README.md
Browse files
README.md
CHANGED
|
@@ -167,57 +167,57 @@ lm_eval \
|
|
| 167 |
<td rowspan="7"><b>OpenLLM V1</b></td>
|
| 168 |
<td>ARC-Challenge (Acc-Norm, 25-shot)</td>
|
| 169 |
<td>37.20</td>
|
| 170 |
-
<td
|
| 171 |
-
<td
|
| 172 |
</tr>
|
| 173 |
<tr>
|
| 174 |
<td>GSM8K (Strict-Match, 5-shot)</td>
|
| 175 |
<td>69.98</td>
|
| 176 |
-
<td
|
| 177 |
-
<td
|
| 178 |
</tr>
|
| 179 |
<tr>
|
| 180 |
<td>HellaSwag (Acc-Norm, 10-shot)</td>
|
| 181 |
<td>43.86</td>
|
| 182 |
-
<td
|
| 183 |
-
<td
|
| 184 |
</tr>
|
| 185 |
<tr>
|
| 186 |
<td>MMLU (Acc, 5-shot)</td>
|
| 187 |
<td>37.38</td>
|
| 188 |
-
<td
|
| 189 |
-
<td
|
| 190 |
</tr>
|
| 191 |
<tr>
|
| 192 |
<td>TruthfulQA (MC2, 0-shot)</td>
|
| 193 |
<td>45.21</td>
|
| 194 |
-
<td
|
| 195 |
-
<td
|
| 196 |
</tr>
|
| 197 |
<tr>
|
| 198 |
<td>Winogrande (Acc, 5-shot)</td>
|
| 199 |
<td>54.30</td>
|
| 200 |
-
<td
|
| 201 |
-
<td
|
| 202 |
</tr>
|
| 203 |
<tr>
|
| 204 |
<td><b>Average Score</b></td>
|
| 205 |
<td><b>47.99</b></td>
|
| 206 |
-
<td><b
|
| 207 |
-
<td><b
|
| 208 |
</tr>
|
| 209 |
<tr>
|
| 210 |
<td rowspan="7"><b>OpenLLM V2</b></td>
|
| 211 |
<td>IFEval (Inst Level Strict Acc, 0-shot)</td>
|
| 212 |
<td>34.37</td>
|
| 213 |
-
<td
|
| 214 |
-
<td
|
| 215 |
</tr>
|
| 216 |
<tr>
|
| 217 |
<td>BBH (Acc-Norm, 3-shot)</td>
|
| 218 |
<td>34.44</td>
|
| 219 |
-
<td
|
| 220 |
-
<td
|
| 221 |
</tr>
|
| 222 |
<tr>
|
| 223 |
<td>Math-Hard (Exact-Match, 4-shot)</td>
|
|
@@ -228,26 +228,26 @@ lm_eval \
|
|
| 228 |
<tr>
|
| 229 |
<td>GPQA (Acc-Norm, 0-shot)</td>
|
| 230 |
<td>24.67</td>
|
| 231 |
-
<td
|
| 232 |
-
<td
|
| 233 |
</tr>
|
| 234 |
<tr>
|
| 235 |
<td>MUSR (Acc-Norm, 0-shot)</td>
|
| 236 |
<td>35.82</td>
|
| 237 |
-
<td
|
| 238 |
-
<td
|
| 239 |
</tr>
|
| 240 |
<tr>
|
| 241 |
<td>MMLU-Pro (Acc, 5-shot)</td>
|
| 242 |
<td>11.80</td>
|
| 243 |
-
<td
|
| 244 |
-
<td
|
| 245 |
</tr>
|
| 246 |
<tr>
|
| 247 |
<td><b>Average Score</b></td>
|
| 248 |
<td><b>23.52</b></td>
|
| 249 |
-
<td><b
|
| 250 |
-
<td><b
|
| 251 |
</tr>
|
| 252 |
<tr>
|
| 253 |
<td rowspan="4"><b>Coding</b></td>
|
|
|
|
| 167 |
<td rowspan="7"><b>OpenLLM V1</b></td>
|
| 168 |
<td>ARC-Challenge (Acc-Norm, 25-shot)</td>
|
| 169 |
<td>37.20</td>
|
| 170 |
+
<td>35.84</td>
|
| 171 |
+
<td>96.3%</td>
|
| 172 |
</tr>
|
| 173 |
<tr>
|
| 174 |
<td>GSM8K (Strict-Match, 5-shot)</td>
|
| 175 |
<td>69.98</td>
|
| 176 |
+
<td>68.01</td>
|
| 177 |
+
<td>97.2%</td>
|
| 178 |
</tr>
|
| 179 |
<tr>
|
| 180 |
<td>HellaSwag (Acc-Norm, 10-shot)</td>
|
| 181 |
<td>43.86</td>
|
| 182 |
+
<td>42.38</td>
|
| 183 |
+
<td>96.6%</td>
|
| 184 |
</tr>
|
| 185 |
<tr>
|
| 186 |
<td>MMLU (Acc, 5-shot)</td>
|
| 187 |
<td>37.38</td>
|
| 188 |
+
<td>36.98</td>
|
| 189 |
+
<td>98.9%</td>
|
| 190 |
</tr>
|
| 191 |
<tr>
|
| 192 |
<td>TruthfulQA (MC2, 0-shot)</td>
|
| 193 |
<td>45.21</td>
|
| 194 |
+
<td>46.68</td>
|
| 195 |
+
<td>103.3%</td>
|
| 196 |
</tr>
|
| 197 |
<tr>
|
| 198 |
<td>Winogrande (Acc, 5-shot)</td>
|
| 199 |
<td>54.30</td>
|
| 200 |
+
<td>55.49</td>
|
| 201 |
+
<td>102.2%</td>
|
| 202 |
</tr>
|
| 203 |
<tr>
|
| 204 |
<td><b>Average Score</b></td>
|
| 205 |
<td><b>47.99</b></td>
|
| 206 |
+
<td><b>47.56</b></td>
|
| 207 |
+
<td><b>99.1%</b></td>
|
| 208 |
</tr>
|
| 209 |
<tr>
|
| 210 |
<td rowspan="7"><b>OpenLLM V2</b></td>
|
| 211 |
<td>IFEval (Inst Level Strict Acc, 0-shot)</td>
|
| 212 |
<td>34.37</td>
|
| 213 |
+
<td>34.42</td>
|
| 214 |
+
<td>100.2%</td>
|
| 215 |
</tr>
|
| 216 |
<tr>
|
| 217 |
<td>BBH (Acc-Norm, 3-shot)</td>
|
| 218 |
<td>34.44</td>
|
| 219 |
+
<td>36.48</td>
|
| 220 |
+
<td>105.9%</td>
|
| 221 |
</tr>
|
| 222 |
<tr>
|
| 223 |
<td>Math-Hard (Exact-Match, 4-shot)</td>
|
|
|
|
| 228 |
<tr>
|
| 229 |
<td>GPQA (Acc-Norm, 0-shot)</td>
|
| 230 |
<td>24.67</td>
|
| 231 |
+
<td>24.78</td>
|
| 232 |
+
<td>100.5%</td>
|
| 233 |
</tr>
|
| 234 |
<tr>
|
| 235 |
<td>MUSR (Acc-Norm, 0-shot)</td>
|
| 236 |
<td>35.82</td>
|
| 237 |
+
<td>35.55</td>
|
| 238 |
+
<td>99.3%</td>
|
| 239 |
</tr>
|
| 240 |
<tr>
|
| 241 |
<td>MMLU-Pro (Acc, 5-shot)</td>
|
| 242 |
<td>11.80</td>
|
| 243 |
+
<td>11.40</td>
|
| 244 |
+
<td>96.6%</td>
|
| 245 |
</tr>
|
| 246 |
<tr>
|
| 247 |
<td><b>Average Score</b></td>
|
| 248 |
<td><b>23.52</b></td>
|
| 249 |
+
<td><b>23.77</b></td>
|
| 250 |
+
<td><b>101.1%</b></td>
|
| 251 |
</tr>
|
| 252 |
<tr>
|
| 253 |
<td rowspan="4"><b>Coding</b></td>
|