DPO Finetuned paulml/OGNO-7B using jondurbin/truthy-dpo-v0.1
paulml/OGNO-7B is a mistral 7b variant afaik and this repo is an experimental repo, so might not be useable in prod
Thx for the great data sources.
Open LLM Leaderboard Evaluation Results
Detailed results can be found here
| Metric | Value |
|---|---|
| Avg. | 76.14 |
| AI2 Reasoning Challenge (25-Shot) | 72.95 |
| HellaSwag (10-Shot) | 89.02 |
| MMLU (5-Shot) | 64.61 |
| TruthfulQA (0-shot) | 76.61 |
| Winogrande (5-shot) | 84.69 |
| GSM8k (5-shot) | 68.99 |
- Downloads last month
- 3,685
Dataset used to train eren23/OGNO-7b-dpo-truthful
Evaluation results
- normalized accuracy on AI2 Reasoning Challenge (25-Shot)test set Open LLM Leaderboard72.950
- normalized accuracy on HellaSwag (10-Shot)validation set Open LLM Leaderboard89.020
- accuracy on MMLU (5-Shot)test set Open LLM Leaderboard64.610
- mc2 on TruthfulQA (0-shot)validation set Open LLM Leaderboard76.610
- accuracy on Winogrande (5-shot)validation set Open LLM Leaderboard84.690
- accuracy on GSM8k (5-shot)test set Open LLM Leaderboard68.990