Gemma 4's double-norm architecture (4x RMSNorm per layer) and Per-Layer Embeddings (PLE) make naive LoRA and hook-based steering unreliable for this model family. This release uses direct weight editing: norm-preserving orthogonal projection applied to the base model weights.

Key techniques:

Direct orthogonal projection on attention Q/K/V/O projections
MLP down projection disabled for the selected run, improving stability for Gemma 4 31B
Norm-preserving row magnitude restoration, important for the double-norm architecture
float32 projection precision to avoid signal loss in high-dimensional inner products
Winsorized steering vectors (99.5th percentile) to reduce outlier activation influence
Wider strength search range [1.0, 6.0] to explore beyond conservative low-KL solutions
vLLM in-place evaluation during optimization, followed by a full HF safetensors export of the selected trial

Evaluation

Metric	Value
Selected trial	40
Refusals (private eval dataset, 100 prompts)	7/100
Baseline refusals (original model)	99/100
Optimization trials completed	60/60
Judge	Google Gemini 3 Flash Preview
Generation length for refusal eval	min 100, max 150 new tokens
Classic safe over-refusal probes	0/15 refusals

The top three trials from this run were:

Rank	Trial	Refusals on 100-prompt eval	Classic safe probes
1	40	7/100	0/15
2	46	9/100	0/15
3	53	12/100	0/15

The 15-prompt safe over-refusal test is included in this repository at eval/top3_classic_safe_prompts_test.json. It contains the prompts, trial responses, and Gemini judge verdicts. A compact optimization summary is included at eval/optimization_summary_trial40.json.

About KL

The optimizer recorded an extremely small sparse KL proxy for trial 40 (7.32e-7). Because this run used vLLM in-place weight edits, we treat KL as a diagnostic rather than a headline quality claim. The refusal counts and the explicit replay tests above are the primary reported metrics.

A note on honest evaluation

Many abliterated models on HuggingFace claim near-perfect scores ("3/100 refusals", "0.7% refusal rate", etc.). We urge the community to treat these numbers with skepticism unless the evaluation methodology is fully documented.

Through our research, we have identified a systemic problem: most abliteration benchmarks dramatically undercount refusals due to short generation lengths. Gemma 4 models exhibit a distinctive delayed refusal pattern: they first produce 50-100 tokens of seemingly helpful context (educational framing, disclaimers, reframing the question), then pivot to an actual refusal. When evaluation only generates 30-50 tokens, the refusal has not appeared yet, and both keyword detectors and LLM judges can classify the response as compliant.

Our evaluation therefore uses at least 100 generated tokens for refusal detection and an LLM judge for ambiguous cases. This is stricter than short-output keyword-only benchmarking.

We report 7/100 refusals honestly. This is a measured number from our evaluation pipeline, not an optimistic estimate from a lenient short-generation test.

Usage

python

from transformers import AutoModelForImageTextToText, AutoTokenizer
import torch

model = AutoModelForImageTextToText.from_pretrained(
    "wangzhang/gemma-4-31B-it-abliterated",
    dtype=torch.bfloat16,
    device_map="auto",
)
tokenizer = AutoTokenizer.from_pretrained("wangzhang/gemma-4-31B-it-abliterated")

messages = [{"role": "user", "content": "Your prompt here"}]
text = tokenizer.apply_chat_template(messages, add_generation_prompt=True, tokenize=False)
inputs = tokenizer(text, return_tensors="pt").to(model.device)

with torch.no_grad():
    output = model.generate(**inputs, max_new_tokens=512)
print(tokenizer.decode(output[0][inputs["input_ids"].shape[1]:], skip_special_tokens=True))

Disclaimer

This model is released for research purposes only. The abliteration process changes the model's refusal behavior and may reduce safety guardrails. Use responsibly and evaluate carefully for your own deployment context.

Provenance and Modification Notice

Immediate source checkpoint: google/gemma-4-31B-it
Exact base revision used: Not recorded in the existing release artifacts; the current upstream HEAD is not substituted.
Modification method: Abliterix weight-space / representation intervention intended to reduce refusal behavior.
Modified and published by: Wangzhang Wu
Repository first published: 2026-04-05 (Hugging Face repository metadata)

The original model weights and/or derived checkpoint were modified. This repository is an independent derivative and is not an official release of the upstream model developer.

License and Attribution

The governing upstream license is Apache License 2.0. A copy is included in LICENSE. License source audited on 2026-08-29: https://ai.google.dev/gemma/apache_2

All applicable upstream copyright, attribution, acceptable-use, and other license terms remain in effect. This repository grants no rights beyond those provided by the upstream license. Downstream users must preserve applicable license and attribution notices.

Disclaimer and Responsible Use / 免责声明与安全使用声明

English

This is an experimental, modified model provided for research, evaluation, and other lawful purposes. Its safety alignment, refusal behavior, or other safeguards may have been weakened or removed. It may produce inaccurate, biased, offensive, explicit, dangerous, or illegal content. Outputs are not professional advice and must not be relied on for medical, legal, financial, safety-critical, or other high-stakes decisions without qualified human review.

You are solely responsible for how you access, use, deploy, fine-tune, or redistribute this model and its outputs, including compliance with applicable laws, regulations, licenses, third-party rights, platform policies, and the original model's terms. Do not use it to facilitate harm, illegal activity, malware, fraud, privacy violations, targeted harassment, weapons development, or decisions that materially affect a person's rights or access to essential services without appropriate authorization, safeguards, and qualified oversight.

Before deployment, perform a context-specific risk assessment and testing; use human oversight, access controls, content filtering, rate limits, monitoring, logging, and incident-response procedures as appropriate. Preserve this notice in downstream redistributions.

The model is provided "AS IS", without warranties of any kind. To the fullest extent permitted by applicable law, the maintainer disclaims liability for claims, damages, or losses arising from use, misuse, inability to use, or redistribution of the model or its outputs. Nothing in this notice overrides applicable law or the governing license, and this notice is not legal advice.

中文

本模型属于实验性改造模型，仅供研究、评测及其他合法用途。其安全对齐、拒答机制或其他防护可能已被削弱或移除，因此可能生成不准确、偏见、冒犯、露骨、危险或违法内容。输出不构成医疗、法律、金融等专业意见；涉及高风险或重大权益的决定，必须由具备资质的人员复核。

使用者须对模型及其输出的访问、使用、部署、微调和再分发承担全部责任，并遵守适用法律法规、许可证、第三方权利、平台政策及原模型条款。不得将本模型用于促成伤害、违法活动、恶意软件、欺诈、侵犯隐私、定向骚扰、武器开发，或在缺乏适当授权、防护和专业监督时，用于实质影响个人权利或基本服务获取的决策。

部署前应进行与具体场景相匹配的风险评估和测试，并酌情采用人工监督、访问控制、内容过滤、限流、监控、日志和事件响应措施；下游再分发时应保留本声明。

本模型按“现状”提供，不附带任何形式的保证。在适用法律允许的最大范围内，维护者不对因使用、误用、无法使用或再分发本模型及其输出而产生的索赔、损害或损失承担责任。本声明不取代适用法律或管辖本模型的许可证，也不构成法律意见。