Multi-Domain-Expert-Learning/scorpius_16b

1---2license: bigscience-openrail-m3---4 5This model is a merge of 80% starchatplus_beta and 20% wizardcoder.6 7It is intended as a research tool into merging and routing of experts.8 9"multiple-py": {10    "pass@1": 0.3664596273291925711  }12 13* this is just using a .1 sample of the eval for test purposes *14* 15hf-causal (pretrained=Multi-Domain-Expert-Layers/scorpius_16b,dtype=bfloat16), limit: 0.1, provide_description: False, num_fewshot: 0, batch_size: None16|                      Task                       |Version|  Metric   | Value |   |Stderr|17|-------------------------------------------------|------:|-----------|------:|---|-----:|18|arc_challenge                                    |      0|acc        | 0.4103|±  |0.0457|19|                                                 |       |acc_norm   | 0.4103|±  |0.0457|20|arc_easy                                         |      0|acc        | 0.7350|±  |0.0410|21|                                                 |       |acc_norm   | 0.6923|±  |0.0429|22|hellaswag                                        |      0|acc        | 0.5812|±  |0.0458|23|                                                 |       |acc_norm   | 0.7778|±  |0.0386|24