Update README.md
Browse files
README.md
CHANGED
|
@@ -23,8 +23,21 @@ The Cogito LLMs are instruction tuned generative models (text in/text out). All
|
|
| 23 |
- In both standard and reasoning modes, Cogito v1-preview models outperform their size equivalent counterparts on common industry benchmarks.
|
| 24 |
- Each model is trained in over 30 languages and supports a context length of 128k.
|
| 25 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 26 |
For detailed evaluations, please refer to the [Blog Post](https://www.deepcogito.com/research/cogito-v1-preview).
|
| 27 |
|
|
|
|
| 28 |
# Usage
|
| 29 |
Here is a snippet below for usage with Transformers:
|
| 30 |
|
|
|
|
| 23 |
- In both standard and reasoning modes, Cogito v1-preview models outperform their size equivalent counterparts on common industry benchmarks.
|
| 24 |
- Each model is trained in over 30 languages and supports a context length of 128k.
|
| 25 |
|
| 26 |
+
# Evaluations
|
| 27 |
+
We compare our models against the state of the art size equivalent models in direct mode as well as the reasoning mode. For the direct mode, we compare against Llama / Qwen instruct counterparts. For reasoning, we use Deepseek's R1 distilled counterparts / Qwen's QwQ model.
|
| 28 |
+
|
| 29 |
+
<p align="left">
|
| 30 |
+
<img src="images/32b_benchmarks.png" alt="Logo" width="90%">
|
| 31 |
+
</p>
|
| 32 |
+
|
| 33 |
+
**Livebench Global Average:**
|
| 34 |
+
<p align="left">
|
| 35 |
+
<img src="images/livebench_global_average.png" alt="Logo" width="80%">
|
| 36 |
+
</p>
|
| 37 |
+
|
| 38 |
For detailed evaluations, please refer to the [Blog Post](https://www.deepcogito.com/research/cogito-v1-preview).
|
| 39 |
|
| 40 |
+
|
| 41 |
# Usage
|
| 42 |
Here is a snippet below for usage with Transformers:
|
| 43 |
|