Modelpublic

surrogate-base-model/oracle-sft-italian-food-post-hoc-unmixed-fd-targeted

sourceHugging Faceupdated 24d agoView on Hugging Face

0likes42downloads

Model Card

LoRA Adapter for SAE Introspection

This is a LoRA (Low-Rank Adaptation) adapter trained for SAE (Sparse Autoencoder) introspection tasks.

Base Model

—Base Model: surrogate-base-model/sft-italian-food-post-hoc-unmixed-fd-targeted
—Adapter Type: LoRA
—Task: SAE Feature Introspection

Usage

python

from transformers import AutoModelForCausalLM, AutoTokenizer
from peft import PeftModel

# Load base model and tokenizer
base_model = AutoModelForCausalLM.from_pretrained("surrogate-base-model/sft-italian-food-post-hoc-unmixed-fd-targeted")
tokenizer = AutoTokenizer.from_pretrained("surrogate-base-model/sft-italian-food-post-hoc-unmixed-fd-targeted")

# Load LoRA adapter
model = PeftModel.from_pretrained(base_model, "surrogate-base-model/oracle-sft-italian-food-post-hoc-unmixed-fd-targeted")

Training Details

This adapter was trained using the lightweight SAE introspection training script to help the model understand and explain SAE features through activation steering.