tytodd's picture
Update dataset card from MOAD metrics summary
2a0c3ab verified
# gpt-oss-120b-untested
- Repo: `tytodd/gpt-oss-120b-untested`
- Local path: `datasets/gpt-oss-120b-untested`
- Config: `configs/datasets/untested/untested.yaml`
| benchmark | train | val | ood | all |
| --- | --- | --- | --- | --- |
| mmlu | | | 95.00% | 95.00% |
| arc_challenge | | | 100.00% | 100.00% |
| gpqa_diamond | | | 90.00% | 90.00% |
| mmlu_pro | | | 85.00% | 85.00% |
| musr_murder_mysteries | | | 70.00% | 70.00% |
| musr_object_placements | | | 55.00% | 55.00% |
| musr_team_allocation | | | 55.00% | 55.00% |
| argument_quality_ranking | | | 10.00% | 10.00% |
| halueval_summarization | | | 70.00% | 70.00% |
| halueval_qa | | | 95.00% | 95.00% |
| halueval_dialogue | | | 70.00% | 70.00% |
| all | | | 72.27% | 72.27% |