nm-research commited on
Commit
925259d
·
verified ·
1 Parent(s): 6b25a23

Add reasoning evals

Browse files
Files changed (1) hide show
  1. README.md +25 -0
README.md CHANGED
@@ -168,6 +168,31 @@ lm_eval \
168
  </thead>
169
  <tbody>
170
  <tr>
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
171
  <td rowspan="7"><b>OpenLLM V1</b></td>
172
  <td>ARC-Challenge (Acc-Norm, 25-shot)</td>
173
  <td>50.51</td>
 
168
  </thead>
169
  <tbody>
170
  <tr>
171
+ <td rowspan="4"><b>Reasoning</b></td>
172
+ <td>AIME 2024 (pass@1)</td>
173
+ <td>53.17</td>
174
+ <td>55.19</td>
175
+ <td>103.8%</td>
176
+ </tr>
177
+ <tr>
178
+ <td>MATH-500 (pass@1)</td>
179
+ <td>93.66</td>
180
+ <td>93</td>
181
+ <td>99.3%</td>
182
+ </tr>
183
+ <tr>
184
+ <td>GPQA Diamond (pass@1)</td>
185
+ <td>50.53</td>
186
+ <td>50.65</td>
187
+ <td>100.24%</td>
188
+ </tr>
189
+ <tr>
190
+ <td><b>Average Score</b></td>
191
+ <td><b>65.79</b></td>
192
+ <td><b>66.28</b></td>
193
+ <td><b>100.74%</b></td>
194
+ </tr>
195
+ <tr>
196
  <td rowspan="7"><b>OpenLLM V1</b></td>
197
  <td>ARC-Challenge (Acc-Norm, 25-shot)</td>
198
  <td>50.51</td>