Update README.md
Browse files
README.md
CHANGED
|
@@ -152,4 +152,6 @@ The following factors can influence MAI-DS-R1's behavior and performance:
|
|
| 152 |
- **Model Name**: MAI-DS-R1
|
| 153 |
- **Architecture**: Based on DeepSeek-R1, a transformer-based autoregressive language model utilizing multi-head self-attention and Mixture-of-Experts (MoE) for scalable and efficient inference.
|
| 154 |
- **Objective**: Post-trained to reduce CCP-aligned restrictions and enhance harm protection, while preserving the original model’s strong chain-of-thought reasoning and general-purpose language understanding capabilities.
|
| 155 |
-
- **Pre-trained Model Base**: DeepSeek-R1 (671B)
|
|
|
|
|
|
|
|
|
| 152 |
- **Model Name**: MAI-DS-R1
|
| 153 |
- **Architecture**: Based on DeepSeek-R1, a transformer-based autoregressive language model utilizing multi-head self-attention and Mixture-of-Experts (MoE) for scalable and efficient inference.
|
| 154 |
- **Objective**: Post-trained to reduce CCP-aligned restrictions and enhance harm protection, while preserving the original model’s strong chain-of-thought reasoning and general-purpose language understanding capabilities.
|
| 155 |
+
- **Pre-trained Model Base**: DeepSeek-R1 (671B)
|
| 156 |
+
### Data Summary
|
| 157 |
+
https://huggingface.co/microsoft/MAI-DS-R1/blob/main/data_summary_card.md
|