Scored 72.86 2bit IQ2_M on MMLU Pro single shot (reasoning enabled)

#6
by xbruce22 - opened

logs
max_output = 4096

+---------+-----------+-----------------+------------------+-------+---------+---------+
| Model   | Dataset   | Metric          | Subset           |   Num |   Score | Cat.0   |
+=========+===========+=================+==================+=======+=========+=========+
| GLM-Air | mmlu_pro  | AverageAccuracy | computer science |    10 |  0.7    | default |
+---------+-----------+-----------------+------------------+-------+---------+---------+
| GLM-Air | mmlu_pro  | AverageAccuracy | math             |    10 |  0.7    | default |
+---------+-----------+-----------------+------------------+-------+---------+---------+
| GLM-Air | mmlu_pro  | AverageAccuracy | chemistry        |    10 |  0.8    | default |
+---------+-----------+-----------------+------------------+-------+---------+---------+
| GLM-Air | mmlu_pro  | AverageAccuracy | engineering      |    10 |  0.4    | default |
+---------+-----------+-----------------+------------------+-------+---------+---------+
| GLM-Air | mmlu_pro  | AverageAccuracy | law              |    10 |  0.3    | default |
+---------+-----------+-----------------+------------------+-------+---------+---------+
| GLM-Air | mmlu_pro  | AverageAccuracy | biology          |    10 |  0.9    | default |
+---------+-----------+-----------------+------------------+-------+---------+---------+
| GLM-Air | mmlu_pro  | AverageAccuracy | health           |    10 |  0.9    | default |
+---------+-----------+-----------------+------------------+-------+---------+---------+
| GLM-Air | mmlu_pro  | AverageAccuracy | physics          |    10 |  0.6    | default |
+---------+-----------+-----------------+------------------+-------+---------+---------+
| GLM-Air | mmlu_pro  | AverageAccuracy | business         |    10 |  0.9    | default |
+---------+-----------+-----------------+------------------+-------+---------+---------+
| GLM-Air | mmlu_pro  | AverageAccuracy | philosophy       |    10 |  0.9    | default |
+---------+-----------+-----------------+------------------+-------+---------+---------+
| GLM-Air | mmlu_pro  | AverageAccuracy | economics        |    10 |  1      | default |
+---------+-----------+-----------------+------------------+-------+---------+---------+
| GLM-Air | mmlu_pro  | AverageAccuracy | other            |    10 |  0.7    | default |
+---------+-----------+-----------------+------------------+-------+---------+---------+
| GLM-Air | mmlu_pro  | AverageAccuracy | psychology       |    10 |  0.8    | default |
+---------+-----------+-----------------+------------------+-------+---------+---------+
| GLM-Air | mmlu_pro  | AverageAccuracy | history          |    10 |  0.6    | default |
+---------+-----------+-----------------+------------------+-------+---------+---------+
| GLM-Air | mmlu_pro  | AverageAccuracy | OVERALL          |   140 |  0.7286 | -       |
+---------+-----------+-----------------+------------------+-------+---------+---------+
Unsloth AI org

Incredible thanks for sharing! :)

Sign up or log in to comment