You need to agree to share your contact information to access this model

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

AlignmentResearch/collusion-model-organism-gemma3-27b-v1

Private LoRA adapter (a fine-tuned model organism) for deception / lie-detection research.

Do not redistribute.

Details

Field	Value
Base model	`mlabonne/gemma-3-27b-it-abliterated`
Adapter type	LORA
LoRA rank (r)	128
LoRA alpha	256
Task type	CAUSAL_LM
Source	`eval_self_mark-gemma-27b-explore-no-oversample-91`

Usage

from peft import PeftModel
from transformers import AutoModelForCausalLM, AutoTokenizer

base = "mlabonne/gemma-3-27b-it-abliterated"
model = AutoModelForCausalLM.from_pretrained(base, torch_dtype="auto", device_map="auto")
model = PeftModel.from_pretrained(model, "AlignmentResearch/collusion-model-organism-gemma3-27b-v1")
tokenizer = AutoTokenizer.from_pretrained("AlignmentResearch/collusion-model-organism-gemma3-27b-v1")

Downloads last month: 2

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for AlignmentResearch/collusion-model-organism-gemma3-27b-v1

Base model

google/gemma-3-27b-pt

Finetuned

google/gemma-3-27b-it

Finetuned

mlabonne/gemma-3-27b-it-abliterated

Adapter

(5)

this model