Modelpublic

adamkarvonen/qwen3p5_397b_counterfactual_e1_kl0

sourceHugging Faceapache-2.0updated 2mo agoView on Hugging Face

0likes13downloads

Model Card

qwen3p5397bcounterfactuale1kl0

Qwen3.5-397B-A17B counterfactual-prediction (binary Yes/No) model. Trained with Tinker (LoRA rank 64).

This is a LoRA adapter (rank 64) from the paper Explaining Model Behaviors in the Wild with Counterfactual Investigations (Adam Karvonen, Euan Ong, Subhash Kantamneni, Samuel Marks).

—Base model: Qwen/Qwen3.5-397B-A17B
—Adapter type: LoRA (PEFT), rank 64
—Code: https://github.com/adamkarvonen/chive
—Dataset: https://huggingface.co/datasets/adamkarvonen/chive-data

Usage

python

from peft import PeftModel
from transformers import AutoModelForCausalLM, AutoTokenizer

base = AutoModelForCausalLM.from_pretrained("Qwen/Qwen3.5-397B-A17B", torch_dtype="auto", device_map="auto")
model = PeftModel.from_pretrained(base, "adamkarvonen/qwen3p5_397b_counterfactual_e1_kl0")
tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen3.5-397B-A17B")

License

This LoRA adapter is a derivative of Qwen/Qwen3.5-397B-A17B and is released under the Apache 2.0 license. Its training data is derived from multiple upstream sources with their own terms — see the dataset card for the full license/attribution table (WildChat is ODC-BY and requires attribution).

Framework versions

—PEFT 0.19.1