Downloads · 30 days
3
3% of all-time downloads
legal-embedding-surgery-part2
SandraChaar1/legal-embedding-surgery-part2
legal-embedding-surgery-part2 is a machine learning model from SandraChaar1. Use it for the machine learning task on the model card, and read the license before you ship it in a product.
- nomic-ai/nomic-embed-text-v1-unsupervised
All-time downloads
96
Public
Parameters
137M
547 MB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors547 MB · 100%
At a glance
- Model type
- nomic_bert
- Access
- Public
- Created
- Jul 18, 2026
- Updated
- Jul 20, 2026
- SHA
- cc25135b
Model card
From the Hugging Face model README
Legal Embedding Surgery – Part 2
Base Model
nomic-ai/nomic-embed-text-v1-unsupervised
Domain
- Jurisprudence
Added Tokens
- tortfeasor
- estoppel
- subpoena
- habeas
- amicus
- interpleader
- judicata
- voir
Implementation
This repository contains a manually modified version of the Nomic embedding model.
The implementation:
- Manually adds new legal tokens to the main WordPiece vocabulary (not
added_tokens). - Does not use
tokenizer.add_tokens(). - Does not use
model.resize_token_embeddings(). - Manually expands the embedding matrix while preserving the original embeddings.
- Initializes embeddings for the newly added tokens.
- Saves and successfully reloads the modified tokenizer and model.
- Verifies that the new tokens are recognized correctly after reloading.
Files
main.py— implementation code- Modified tokenizer
- Modified model weights
- Type
- nomic_bert
- Created
- Jul 18, 2026
- Updated
- Jul 20, 2026