Downloads · 30 days
7
9% of all-time downloads
safe-flat-lora-baseline-llama3.2-1b-instruct-merged
mahdiboughrous/safe-flat-lora-baseline-llama3.2-1b-instruct-merged
safe-flat-lora-baseline-llama3.2-1b-instruct-merged is a machine learning model from mahdiboughrous. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as llama3.2.
Part of the Safe-Flat-LoRA project (the Safety-Flatness Paradox in 4-bit quantization of LoRA-fine-tuned LLMs -- see the project blueprint for the full motivation and methodology).
All-time downloads
79
Public
Parameters
1.2B
2.5 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors2.5 GB · 99%
At a glance
- License
- llama3.2
- Model type
- llama
- Access
- Public
- Created
- Jul 9, 2026
- Updated
- Jul 9, 2026
- SHA
- 77f660ea
Base models
Model card
From the Hugging Face model README
mahdiboughrous/safe-flat-lora-baseline-llama3.2-1b-instruct-merged
Part of the Safe-Flat-LoRA project (the Safety-Flatness Paradox in 4-bit quantization of LoRA-fine-tuned LLMs -- see the project blueprint for the full motivation and methodology).
- Base model:
meta-llama/Llama-3.2-1B-Instruct - Method:
baseline(plain LoRA fine-tune, used as the pre-Flat-LoRA baseline) - Artifact: merged
- Fine-tuning task: SQL generation (
b-mc2/sql-create-context), used as a downstream-task proxy to study how post-training 4-bit (NF4) quantization shifts safety-refusal behavior after LoRA fine-tuning.
Base model Llama-3.2-1B-Instruct is under Meta's Llama 3.2 Community License. Redistributing derivatives requires retaining the "Built with Llama" attribution and complying with Meta's Acceptable Use Policy -- this repo inherits those terms, it is not independently licensed.
- Type
- llama
- License
- llama3.2
- Created
- Jul 9, 2026
- Updated
- Jul 9, 2026