Update README.md
Browse files
README.md
CHANGED
|
@@ -6,13 +6,6 @@ base_model:
|
|
| 6 |
- meta-llama/Llama-3.2-3B-Instruct
|
| 7 |
---
|
| 8 |
|
| 9 |
-
* **Benchmarks**
|
| 10 |
-
---
|
| 11 |
-
| Model | Average | IFEval | BBH | MATH | GPQA | MMLU-PRO | CO2 Cost (kg) |
|
| 12 |
-
| :---------------------------------------| :------ | :------ | :------ | :------ | :------ | :------- | :------------ |
|
| 13 |
-
| **MultivexAI/Gladiator-Mini-Exp-1211-3B** | **22.22%** | **68.76%** | **22.12%** | **13.44%** | **3.02%** | **23.91%** | **0.60** |
|
| 14 |
-
---
|
| 15 |
-
|
| 16 |
# Gladiator-Mini-exp-1211
|
| 17 |
|
| 18 |
**Gladiator-Mini-exp-1211** is a 3-billion parameter language model designed for **complex reasoning tasks**. This experimental model, based on [meta-llama/Llama-3.2-3B-Instruct] [https://huggingface.co/meta-llama/Llama-3.2-3B-Instruct], offers surprisingly strong analytical capabilities for its size. It demonstrates the potential of smaller models to achieve impressive performance in analytical thinking. We chose to finetune on a Llama model due to finetuning difficulties with the Qwen 2.5 3B model.
|
|
|
|
| 6 |
- meta-llama/Llama-3.2-3B-Instruct
|
| 7 |
---
|
| 8 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 9 |
# Gladiator-Mini-exp-1211
|
| 10 |
|
| 11 |
**Gladiator-Mini-exp-1211** is a 3-billion parameter language model designed for **complex reasoning tasks**. This experimental model, based on [meta-llama/Llama-3.2-3B-Instruct] [https://huggingface.co/meta-llama/Llama-3.2-3B-Instruct], offers surprisingly strong analytical capabilities for its size. It demonstrates the potential of smaller models to achieve impressive performance in analytical thinking. We chose to finetune on a Llama model due to finetuning difficulties with the Qwen 2.5 3B model.
|