| \# SNP-Universal-Embedding | |
| \_A public-ready embedding model derived from the Substrate-Prism Neuron (SNP) framework.\_ | |
| \## Description | |
| SNP-Universal-Embedding is a 6-dimensional embedding model built from emotional geometry principles | |
| originating in the Substrate-Prism Neuron research. | |
| It maps semantic and emotional relationships into compact geometric space — designed for | |
| research in cognitive modeling, decision conflict analysis, and affective reasoning. | |
| \## Key Info | |
| \- \*\*Base model:\*\* BERT-base-uncased | |
| \- \*\*Dimensions:\*\* 6 | |
| \- \*\*Purpose:\*\* Compact universal embeddings reflecting emotional \& relational context | |
| \- \*\*Use cases:\*\* | |
| - Compare semantic + affective similarity | |
| - Feed into downstream emotional-reasoning models | |
| - Multi-modal integration (text → cognitive vector space) | |
| \## Example Usage (local) | |
| ```python | |
| from transformers import AutoTokenizer, AutoModel | |
| import torch | |
| tokenizer = AutoTokenizer.from\_pretrained("./SNP-Universal-Embedding") | |
| model = AutoModel.from\_pretrained("./SNP-Universal-Embedding") | |
| inputs = tokenizer("A decision between love and duty.", return\_tensors="pt") | |
| with torch.no\_grad(): | |
| output = model(\*\*inputs) | |
| embedding = output.last\_hidden\_state.mean(dim=1) | |
| print(embedding.shape) # torch.Size(\[1, 6]) | |