| Model Type | | translation, comprehension |
|
| Use Cases |
| Areas: | | Research, Commercial applications |
|
| Applications: | | Translation, Comprehension |
|
|
| Additional Notes | | Future plans include SLERP merging and potentially creating larger models using MergeKit. |
|
| Supported Languages | | Hebrew (understands and comprehends), English (decents translation) |
|
| Training Details |
| Methodology: | | used SOTA techniques and personal insights |
|
| Model Architecture: | | Based on Mistral 7B, tokenizer not altered. |
|
|
| Input Output |
| Input Format: | | Debug-deterministic with specified parameters. |
|
| Accepted Modalities: | |
| Performance Tips: | | No specific tips provided beyond using the recommended prompt settings. |
|
|
| Release Notes |
| Version: | |
| Date: | |
| Notes: | | Acquired the highest SNLI score among open source models in Hebrew. |
|
|
|