Update README.md
Browse files
README.md
CHANGED
|
@@ -63,19 +63,6 @@ In AlpacaEval, Rocket 🦝 achieves a near 80% win rate, coupled with an average
|
|
| 63 |
| **Rocket** 🦝 | **79.75** | **1.42** | **1242** |
|
| 64 |
|
| 65 |
|
| 66 |
-
## Open LLM leaderboard
|
| 67 |
-
|
| 68 |
-
| Metric | Value |
|
| 69 |
-
|-----------------------|---------------------------|
|
| 70 |
-
| Average | 55.77 |
|
| 71 |
-
| ARC | 50.6 |
|
| 72 |
-
| HellaSwag | 76.69 |
|
| 73 |
-
| MMLU | 47.1 |
|
| 74 |
-
| TruthfulQA | 55.82 |
|
| 75 |
-
| Winogrande | 67.96 |
|
| 76 |
-
| GSM8K | 36.47 |
|
| 77 |
-
|
| 78 |
-
|
| 79 |
## Intended uses & limitations
|
| 80 |
Initially, we fine-tuned the model using a dataset created by merging and curating multiple datasets, available on the HuggingFace Hub. This dataset will be released to the public soon. We further enhanced the model's performance using DPO, selecting samples from the [openbmb/UltraFeedback](https://huggingface.co/datasets/openbmb/UltraFeedback) and [BAAI/JudgeLM-100K](https://huggingface.co/datasets/BAAI/JudgeLM-100K) datasets. The outcome is a highly effective chat model with a 3 billion parameter scale.
|
| 81 |
|
|
|
|
| 63 |
| **Rocket** 🦝 | **79.75** | **1.42** | **1242** |
|
| 64 |
|
| 65 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 66 |
## Intended uses & limitations
|
| 67 |
Initially, we fine-tuned the model using a dataset created by merging and curating multiple datasets, available on the HuggingFace Hub. This dataset will be released to the public soon. We further enhanced the model's performance using DPO, selecting samples from the [openbmb/UltraFeedback](https://huggingface.co/datasets/openbmb/UltraFeedback) and [BAAI/JudgeLM-100K](https://huggingface.co/datasets/BAAI/JudgeLM-100K) datasets. The outcome is a highly effective chat model with a 3 billion parameter scale.
|
| 68 |
|