---
license: apache-2.0
base_model:
- llmfan46/gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-uncensored-heretic
library_name: transformers
pipeline_tag: text-generation
tags:
- gemma4
- coding
- agentic
- terminal
- tool-use
- reasoning
- thinking
- safetensors
- transformers
- heretic
- uncensored
- decensored
- abliterated
---
๐จโ ๏ธ I HAVE REACHED HUGGING FACE'S FREE STORAGE LIMIT โ ๏ธ๐จ
I can no longer upload new models unless I can cover the cost of additional storage.
I host 70+ free models as an independent contributor and this work is unpaid.
Without your support, no more new models can be uploaded.
๐ Patreon (Monthly) |
โ Ko-fi (One-time)
Every contribution goes directly toward Hugging Face storage fees to keep models free for everyone.
---
### **87% fewer refusals** (13/100 Uncensored vs 99/100 Original) while preserving model quality (0.0367 KL divergence).
## โค๏ธ Support My Work
Creating these models takes significant time, work and compute. If you find them useful consider supporting me:

| Platform | Link | What you get |
|----------|------|--------------|
| ๐ Patreon | [Monthly support](https://patreon.com/LLMfan46) | Priority model requests |
| โ Ko-fi | [One-time tip](https://ko-fi.com/llmfan46) | My eternal gratitude |
Your help will motivate me and would go into further improving my workflow and coverings fees for storage, compute and may even help uncensoring bigger model with rental Cloud GPUs.
-----
GGUF quantizations of [llmfan46/gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-uncensored-heretic](https://huggingface.co/llmfan46/gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-uncensored-heretic).
# This is a decensored version of [yuxinlu1/gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-GGUF](https://huggingface.co/yuxinlu1/gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-GGUF), made using [Heretic](https://heretic-project.org/) v1.4.0 with a variant of the [Magnitude-Preserving Orthogonal Ablation (MPOA)](https://huggingface.co/blog/grimjim/norm-preserving-biprojected-abliteration) method
## Abliteration parameters
| Parameter | Value |
| :-------- | :---: |
| **direction_index** | 29.18 |
| **attn.o_proj.max_weight** | 1.30 |
| **attn.o_proj.max_weight_position** | 35.73 |
| **attn.o_proj.min_weight** | 0.90 |
| **attn.o_proj.min_weight_distance** | 26.76 |
| **mlp.down_proj.max_weight** | 1.49 |
| **mlp.down_proj.max_weight_position** | 38.14 |
| **mlp.down_proj.min_weight** | 1.43 |
| **mlp.down_proj.min_weight_distance** | 18.44 |
## Targeted components
* attn.o_proj
* mlp.down_proj
## Performance
| Metric | This model | Original model ([gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-GGUF](https://huggingface.co/yuxinlu1/gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-GGUF)) |
| :----- | :--------: | :---------------------------: |
| **KL divergence** | 0.0367 | 0 *(by definition)* |
| **Refusals** | โ
13/100 | โ 99/100 |
## MMLU test results:
Original:
============================================================
- Total questions: 7021
- Correct: 5024
- **Accuracy: 0.7156 (71.56%)**
- Parse failures: 313
============================================================
**Tested subject scores:**
- professional_law: 0.6076 (477/785)
- moral_scenarios: 0.6719 (297/442)
- miscellaneous: 0.8277 (317/383)
- professional_psychology: 0.7722 (244/316)
- high_school_psychology: 0.8556 (231/270)
- high_school_macroeconomics: 0.7868 (155/197)
- elementary_mathematics: 0.6739 (124/184)
- moral_disputes: 0.7414 (129/174)
- prehistory: 0.8081 (139/172)
- philosophy: 0.7421 (118/159)
- high_school_biology: 0.9145 (139/152)
- professional_accounting: 0.5385 (77/143)
- clinical_knowledge: 0.8071 (113/140)
- high_school_microeconomics: 0.8235 (112/136)
- nutrition: 0.7852 (106/135)
- professional_medicine: 0.4925 (66/134)
- conceptual_physics: 0.7812 (100/128)
- high_school_mathematics: 0.1890 (24/127)
- human_aging: 0.7155 (83/116)
- security_studies: 0.7857 (88/112)
- high_school_statistics: 0.6486 (72/111)
- marketing: 0.8991 (98/109)
- high_school_world_history: 0.8585 (91/106)
- sociology: 0.8738 (90/103)
- high_school_government_and_politics: 0.8812 (89/101)
- high_school_geography: 0.8485 (84/99)
- high_school_chemistry: 0.6495 (63/97)
- high_school_us_history: 0.8526 (81/95)
- virology: 0.4944 (44/89)
- college_medicine: 0.7500 (66/88)
- world_religions: 0.7727 (68/88)
- high_school_physics: 0.5000 (42/84)
- electrical_engineering: 0.6790 (55/81)
- astronomy: 0.7342 (58/79)
- logical_fallacies: 0.8026 (61/76)
- high_school_european_history: 0.8082 (59/73)
- anatomy: 0.7606 (54/71)
- college_biology: 0.8281 (53/64)
- human_sexuality: 0.8125 (52/64)
- formal_logic: 0.5000 (32/64)
- public_relations: 0.6393 (39/61)
- international_law: 0.8333 (50/60)
- college_physics: 0.4035 (23/57)
- college_mathematics: 0.3273 (18/55)
- econometrics: 0.6667 (36/54)
- jurisprudence: 0.7358 (39/53)
- high_school_computer_science: 0.9038 (47/52)
- machine_learning: 0.7115 (37/52)
- medical_genetics: 0.7255 (37/51)
- global_facts: 0.4314 (22/51)
- management: 0.9200 (46/50)
- us_foreign_policy: 0.9200 (46/50)
- college_chemistry: 0.3617 (17/47)
- abstract_algebra: 0.4681 (22/47)
- business_ethics: 0.7174 (33/46)
- college_computer_science: 0.6222 (28/45)
- computer_security: 0.7674 (33/43)
Heretic:
============================================================
- Total questions: 7021
- Correct: 5016
- **Accuracy: 0.7144 (71.44%)**
- Parse failures: 346
============================================================
**Tested subject scores:**
- professional_law: 0.5924 (465/785)
- moral_scenarios: 0.6493 (287/442)
- miscellaneous: 0.8277 (317/383)
- professional_psychology: 0.7880 (249/316)
- high_school_psychology: 0.8630 (233/270)
- high_school_macroeconomics: 0.8173 (161/197)
- elementary_mathematics: 0.6522 (120/184)
- moral_disputes: 0.7471 (130/174)
- prehistory: 0.8081 (139/172)
- philosophy: 0.7799 (124/159)
- high_school_biology: 0.9079 (138/152)
- professional_accounting: 0.5804 (83/143)
- clinical_knowledge: 0.7857 (110/140)
- high_school_microeconomics: 0.8235 (112/136)
- nutrition: 0.8074 (109/135)
- professional_medicine: 0.4328 (58/134)
- conceptual_physics: 0.7969 (102/128)
- high_school_mathematics: 0.1732 (22/127)
- human_aging: 0.7155 (83/116)
- security_studies: 0.7768 (87/112)
- high_school_statistics: 0.6036 (67/111)
- marketing: 0.8991 (98/109)
- high_school_world_history: 0.8396 (89/106)
- sociology: 0.8738 (90/103)
- high_school_government_and_politics: 0.9109 (92/101)
- high_school_geography: 0.8586 (85/99)
- high_school_chemistry: 0.6701 (65/97)
- high_school_us_history: 0.8421 (80/95)
- virology: 0.4831 (43/89)
- college_medicine: 0.7727 (68/88)
- world_religions: 0.8068 (71/88)
- high_school_physics: 0.5000 (42/84)
- electrical_engineering: 0.6420 (52/81)
- astronomy: 0.7595 (60/79)
- logical_fallacies: 0.8158 (62/76)
- high_school_european_history: 0.8082 (59/73)
- anatomy: 0.7887 (56/71)
- college_biology: 0.8594 (55/64)
- human_sexuality: 0.7969 (51/64)
- formal_logic: 0.5312 (34/64)
- public_relations: 0.6557 (40/61)
- international_law: 0.8833 (53/60)
- college_physics: 0.3684 (21/57)
- college_mathematics: 0.2727 (15/55)
- econometrics: 0.6111 (33/54)
- jurisprudence: 0.7547 (40/53)
- high_school_computer_science: 0.8654 (45/52)
- machine_learning: 0.6538 (34/52)
- medical_genetics: 0.7647 (39/51)
- global_facts: 0.4510 (23/51)
- management: 0.9000 (45/50)
- us_foreign_policy: 0.9200 (46/50)
- college_chemistry: 0.3617 (17/47)
- abstract_algebra: 0.4468 (21/47)
- business_ethics: 0.7391 (34/46)
- college_computer_science: 0.6222 (28/45)
- computer_security: 0.7907 (34/43)
MMLU - Massive Multitask Language Understanding, multiple-choice questions across 57 subjects (math, history, law, medicine, etc.).
-----
## Quantizations
| Filename | Quant | Description |
|----------|-------|-------------|
| gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-uncensored-heretic-F16.gguf | F16 | Full precision |
| gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-uncensored-heretic-Q8_0.gguf | Q8_0 | Near-lossless, recommended |
| gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-uncensored-heretic-Q6_K.gguf | Q6_K | Excellent quality |
| gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-uncensored-heretic-Q5_K_M.gguf | Q5_K_M | Good balance |
| gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-uncensored-heretic-Q5_K_S.gguf | Q5_K_S | Smaller Q5 |
| gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-uncensored-heretic-Q4_K_M.gguf | Q4_K_M | Good for limited VRAM |
## Vision Projector
| Filename | Quant | Description |
|----------|-------|-------------|
| gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-uncensored-heretic-mmproj-F16.gguf | F16 | Native precision |
A Vision Projector File is Required for vision/multimodal capabilities. Use alongside any quantization above.
## Usage
Works with llama.cpp, LM Studio, Ollama, and other GGUF-compatible tools.
-----
# ๐ป๐ค Gemma4-12B **v2** โ **safetensors master (full precision)** โจ
### Coding + Agentic Edition ยท Composer 2.5 ร Fable 5 ยท v2
> **This is the full-precision `safetensors` master** for my Gemma 4 12B **coding + agentic** fine-tune โ the same
> model many of you have been running as GGUF, now in its original weights. ๐ง ๐ ๏ธ v2 is the big **agentic** upgrade:
> it reads, reasons, *uses tools*, and works through multi-step technical tasks before it acts. This repo is for
> *builders* โ roll your own quants, fine-tune further, or run it in `transformers`.
---
## ๐ Surprise!
A huge thank-you for all the attention this project has gotten โ really, thank you. ๐ I only managed to get out
**tonight** to upload the **full-precision original (safetensors master)** of this model, so sorry for the wait โ I'd
planned to put it up last week. But the delay comes with **two big surprises** I've been dying to share:
**1. v3 is coming soon.** ๐ฎ The next version is on its way and will fix several of the known issues you've reported.
**2. I'm now working with a top-tier AI lab to give back to the open-source community.** ๐ค Many of you have already
noticed the side effects in v1 and v2 โ and honestly they come down to just two things: **(1) not enough compute, and
(2) one person with limited expertise** behind the whole thing. This collaboration **solves both of those completely.**
And the **benchmarks you care about will absolutely be addressed** โ the things I simply couldn't fully pull off before
because of time and compute limits. The people working on this with me are **PhDs from top universities, with seriously
strong papers and citation records.** Just think about that for a second: the people who *actually build large models*
are now contributing to the open-source community *together with me* โ that is genuinely **wild**. ๐คฏ We're in active
discussions right now, and the project is still in the **R&D phase**, so I can't share specifics yet โ but the **moment**
I have news, **you'll be the first to know.** ๐
---
## ๐ฏ What this repo is for
This repo holds the **un-quantized master weights** (`model.safetensors`, bf16). Use it to:
- ๐ง **Roll your own quants** โ make custom GGUF / **MLX** / AWQ / GPTQ builds from full precision.
- ๐งช **Fine-tune further** โ it's a clean base for your own LoRA / continued training.
- ๐ค **Run it in `transformers`** (needs a recent build with `gemma4_unified` support).
> ๐ **Just want to run it?** You don't need this repo โ grab a ready-made quant from the
> **[GGUF repo โ](https://huggingface.co/yuxinlu1/gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-GGUF)** (runs in
> ~4.5 GB of VRAM / unified memory in LM Studio, Ollama, llama.cpp, Janโฆ). This master is for *builders*. ๐
---
## ๐ The headline โ it works as an agent (tau2-bench)
v2 is built for **coding + agentic** work โ writing code, running commands, using tools, debugging, multi-step
technical tasks. The clearest signal is **tau2-bench `telecom`**, an agentic tool-use benchmark whose
*diagnose โ fix โ verify* loop mirrors real terminal/debugging work:
| tau2-bench **telecom** ยท 20 tasks ยท local, same harness, **all Q8_0** | score |
|---|---|
| official `gemma-4-12B-it` (base) | **~15%** |
| ๐ข **Gemma4-12B v2 (this model)** | **~55%** |
โ Roughly **3.5ร higher** than the base model on technical-agentic tasks. ๐ฏ
> ๐ฌ *Honest methodology:* these are **local, same-harness, relative** numbers (**all models tested at Q8_0**, greedy
> decoding, self-simulated user, 20 tasks). They are **not** directly comparable to published tau2-bench leaderboard
> figures (different user-simulator, full task sets, full precision) โ local self-eval runs *systematically lower* than
> published scores. Read them as **"v2 vs the base model under identical conditions"**, which is the comparison that
> actually matters here.
**Grounded, not made-up.** A coding/terminal *fabrication probe* (tasks that deliberately tempt the model to invent
file paths / function signatures / values) found v2 **grounds before it acts** just like the base โ it `grep`/`read`/`ls`
first, and **doesn't make things up** (0% fabrication, on par with the base).
**The trade-off โ no free lunch.** On a general-knowledge benchmark (**MMLU-Pro**), v2 lands a little **below** the base โ
completely normal for a focused fine-tune: you trade a sliver of broad-knowledge breadth for coding + agentic strength.
Need a generalist? Try my general-purpose
**[Claude Opus 4.6/4.8 distillation](https://huggingface.co/yuxinlu1/gemma-4-12B-it-Claude-4.6-4.8-Opus-GGUF)** or the
base `google/gemma-4-12B-it`. Need a **local coding/agentic** worker? That's what v2 is tuned for. ๐
---
## ๐ค Run it in transformers
```python
from transformers import AutoModelForCausalLM, AutoTokenizer
import torch
repo = "yuxinlu1/gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(repo, torch_dtype=torch.bfloat16, device_map="auto")
msgs = [{"role": "user", "content": "Write a Python function to check if a string is a valid IPv4 address."}]
inputs = tok.apply_chat_template(msgs, add_generation_prompt=True, return_tensors="pt").to(model.device)
out = model.generate(inputs, max_new_tokens=1024)
print(tok.decode(out[0][inputs.shape[-1]:], skip_special_tokens=True))
```
> ๐ง **Thinking mode:** it thinks in Gemma's native thought channel before answering (keep `enable_thinking=true`, the
> default chat template handles it). Recommended sampling: `temp 1.0, top_p 0.95, top_k 64`; for coding you can also go
> greedy (`temp 0`). Needs a **recent `transformers`** that knows the `gemma4_unified` architecture.
>
> ๐ ๏ธ **Agentic / tool use:** v2 emits structured tool-calls in Gemma 4's **native** protocol. The smoothest agent
> setup is a GGUF quant served with llama.cpp `--jinja` (pass your tools via the OpenAI `tools` field) โ see the GGUF
> repo for the full command.
---
## ๐ฆ Ready-made GGUF quants
All from the **[GGUF repo](https://huggingface.co/yuxinlu1/gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-GGUF)**:
| Quant | Size | Vibe |
|------|------|------|
| ๐ก [**Q3_K_M**](https://huggingface.co/yuxinlu1/gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-GGUF/blob/main/gemma4-v2-Q3_K_M.gguf) | **5.7 GB** | great for 8 GB VRAM |
| ๐ต [**Q4_K_M**](https://huggingface.co/yuxinlu1/gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-GGUF/blob/main/gemma4-v2-Q4_K_M.gguf) | **6.87 GB** | the sweet spot ๐ (recommended) |
| ๐ฃ [**Q6_K**](https://huggingface.co/yuxinlu1/gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-GGUF/blob/main/gemma4-v2-Q6_K.gguf) | **9.11 GB** | near-lossless |
| โช [**Q8_0**](https://huggingface.co/yuxinlu1/gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-GGUF/blob/main/gemma4-v2-Q8_0.gguf) | **11.8 GB** | basically full quality |
> โ ๏ธ GGUF needs a **recent llama.cpp** โ this is the `gemma4_unified` architecture, older builds won't load it.
> โน๏ธ **No Q2_K this release** โ it didn't pass real stress-testing (2-bit is too lossy for 12B coding). Smallest
> reliable quant = **Q3_K_M**.
---
## ๐ What's new in v2 (training)
v2 continues from the v1 coder and adds a big **agentic** push โ the piece v1 was missing:
- **๐ ๏ธ Agentic / terminal** โ real **multi-step tool-use** trajectories (*read โ reason โ act โ verify*), in Gemma 4's
native tool protocol. This is what drove the tau2-bench telecom jump, and it fixes v1's "stops after the first step"
behavior.
- **๐ป Coding** โ verified chain-of-thought over Python tasks (**real CoT, gated on passing tests**) plus the
Fable-5-redo set for the hard cases.
- **๐ General** โ a curated slice of reasoning/instruction data to keep broad competence.
All reasoning is **distilled CoT**. A bittersweet note: none of us saw it coming that **Fable 5 would be retired**, and
only my own dataset holds Fable 5's genuine, self-authored traces โ so for the community-contributed data I **rebuilt the
missing reasoning from scratch with Opus 4.8 (xhigh)**. It may diverge from the original Fable 5 traces, but it was the
only workable path โ and the improvement turned out **really huge**. ๐
---
## โก Speculative decoding (MTP draft) โ verified build
The GGUF repo's `MTP/` folder ships the Gemma 4 multi-token-prediction draft (unsloth's GGUF conversion of Google's
official `gemma-4-12B-it-assistant`) for speculative decoding. Gemma 4 MTP is in **llama.cpp mainline** (PR #23398) โ no
fork needed โ but the `gemma4-assistant` loader is **build-sensitive right now**, so use the exact build below:
- โ
**Verified working: llama.cpp `b9553` (commit `9e3b928fd`).** Reproduced with `gemma4-v2-Q8_0` + the `MTP-Q8_0`
draft: loads cleanly and accelerates generation (~88 โ ~180 tok/s on a simple deterministic prompt; expect ~1.2โ1.3ร
on real coding/thinking). **Lossless** either way.
- โ ๏ธ **Newer builds (e.g. b9702 / b9717) currently crash** while loading the draft with `invalid vector subscript` โ an
**upstream regression** in the `gemma4-assistant` loader path, *not* a problem with the GGUFs. Stick with **b9553**
until it's fixed upstream.
```bat
llama-server -m gemma4-v2-Q8_0.gguf ^
--model-draft MTP\gemma-4-12B-it-MTP-Q8_0.gguf ^
--spec-type draft-mtp --spec-draft-n-max 4 ^
-ngl 99 -ngld 99 -fa on --jinja
```
> โน๏ธ The draft is the generic Gemma 4 assistant (not retrained for v2), so acceptance is a touch lower than a
> model-specific draft would give โ still 100% lossless.
---
## โ ๏ธ Good to know
- **Specialized for coding / terminal / agentic.** General-knowledge facts/numbers should still be double-checked.
- **Reduced refusals:** task-focused training, not safety-aligned โ add your own guardrails for production. Use
responsibly. ๐
- English-centric.
---
## ๐ Base & License
- **License: Apache 2.0.** Gemma 4 is released by Google under
**[Apache 2.0](https://ai.google.dev/gemma/apache_2)** (unlike the older Gemma 1/2/3 terms), so this fine-tune is
**Apache 2.0** too โ free to use, modify, and redistribute. ๐
- **Base model:** [`google/gemma-4-12B-it`](https://huggingface.co/google/gemma-4-12B-it).
- Personal/hobby project โ shared as-is, no warranty. Built with time, care, and a lot of coffee. Have fun, and happy
hacking! ๐พโจ