| language: en | |
| tags: | |
| - hypertensor | |
| - ceci-graft | |
| - danish | |
| - smollm2 | |
| - experimental | |
| pipeline_tag: text-generation | |
| license: apache-2.0 | |
| # minKæreste (my dearest) | |
| Layer 8 mixed-region crossover. FFN transplanted from layer 18. 122% PPL recovery — the strongest graft in the series. PPL drops from 83 (ablated) to 56 (grafted), recovering past the original 61 baseline. The crossed processing pathway creates emergent behavior neither parent layer exhibits alone. | |
| ## Architecture | |
| - **Base**: SmolLM2-135M-Instruct | |
| - **Method**: CECI Protocol (HyperTensor Paper X) — GRC basis projection | |
| - **Created**: 2026-05-04 | |
| - **Repository**: [HyperTensor](https://github.com/NagusameCS/HyperTensor) | |
| ## Graft Proof | |
| This model was created by: | |
| 1. Computing the GRC (Geodesic Residual Compression) basis from the target layer's attention weights via SVD | |
| 2. Projecting the donor layer's FFN weights into the target's geometric subspace | |
| 3. Blending at controlled strength to preserve stability | |
| Perplexity testing confirms the graft transfers functional structure without destroying the model. | |
| ## Usage | |
| ```python | |
| from transformers import AutoModelForCausalLM, AutoTokenizer | |
| model = AutoModelForCausalLM.from_pretrained("NagusameCS/minKæreste", trust_remote_code=True) | |
| tokenizer = AutoTokenizer.from_pretrained("NagusameCS/minKæreste") | |
| ``` | |