allura-org / Model details

Gemma-3-Glitter-27B

allura-org/Gemma-3-Glitter-27B

merged

Read the original model card ↗ · Choosing a model

About this model

The repository allura-org/Gemma-3-Glitter-27B is provided by the creator allura-org and offers a creative‑writing variant of Gemma 3 27B. According to the model card, the model is built on the Columbidae/gemma-3-27b-half checkpoint, which itself is a 50/50 merge of the Google Gemma‑3 27B instruction‑tuned (IT) and pre‑trained (PT) checkpoints, and the creator selected this base because of the success of a similar IT/PT merge in the Starshine model. The creator states that adding the PT component weakens the instruction tuning but also reduces censorship and hesitancy, making the prose feel more natural and better suited for fictional storytelling. The creator notes that the model performs best with short, to‑the‑point prompts and that long, detailed system prompts (1000‑2000 tokens) often lead to lackluster results compared with 100‑500 token prompts. The weights are distributed as safetensors files, and the repository does not provide a license specification.

Description sources

https://huggingface.co/allura-org/Gemma-3-Glitter-27B/raw/c707596609f9a45c7fe3babe30e0d238a92b8364/README.md ↗

Source-based description, not an independent evaluation. Check the linked sources for the creator’s claims and limitations.

Size and formats

Base model: Columbidae/gemma-3-27b-half, google/gemma-3-27b-it, google/gemma-3-27b-pt
Parameters: 27.4324B
Main repository formats: safetensors
License tag: Unknown
Languages: Unknown
Last updated: 2025-04-18 · 03:44:11.000 UTC

What changed

Model types are creator-reported, not independently tested. These labels describe source metadata, not a guarantee of behavior.

mergedsource tag: base_model:merge:Columbidae/gemma-3-27b-half
mergedsource tag: base_model:merge:google/gemma-3-27b-it
mergedsource tag: base_model:merge:google/gemma-3-27b-pt

Family evidence

These groupings are source-reported. Explicit base identifiers take priority over generic architecture tags.

Gemma: Base model: Columbidae/gemma-3-27b-half
Gemma: Base model: google/gemma-3-27b-it
Gemma: Base model: google/gemma-3-27b-pt

Available versions

Versions appear here only when they explicitly name this exact model as their source. Check each repository for available files.

Main repository files ↗safetensors
No other explicitly linked versions recorded.

Link to this listing

Made or quantized this model? Paste one of these into your model card or README so people can find this listing.

Markdown

[Listed on Custom Clankers](https://customclankers.com/models/allura-org-gemma-3-glitter-27b/)

HTML

<a href="https://customclankers.com/models/allura-org-gemma-3-glitter-27b/">Listed on Custom Clankers</a>

Plain text links only: no image, no script, nothing loaded from this site.

Customize or build on this model

Listed from public GitHub data, not tested by us. Read the code and license before you run anything.

Tools to customize it further

Remove more refusals

Heretic ↗ ★ 36,457
Removes refusal behavior from transformer models automatically. It pairs directional ablation (abliteration) with an Optuna parameter search that tries to cut refusals while keeping the model close to the original, so you don't have to pick layers by hand.
OBLITERATUS ↗ ★ 8,949
A larger refusal-removal toolkit with several abliteration methods, usable locally or through a hosted Hugging Face Space. Check the telemetry setting before you run it: the README describes telemetry as opt-in, but on by default in the Spaces version.

Fine-tune it on your own data

Unsloth ↗ ★ 78,549
A desktop app and Python library for running and training models locally. Its README lists LoRA, QLoRA, full fine-tuning, DPO and GRPO, plus export to formats including GGUF, on Windows, Linux, WSL and macOS.
LlamaFactory ↗ ★ 75,720
Fine-tunes more than 100 language and vision-language models from a command line or a browser-based Web UI, aimed at training without writing code.

Make a smaller download (quantize)

llama.cpp ↗ ★ 132,313
Runs models locally and includes the tools behind most GGUF files. A conversion script turns Hugging Face weights into GGUF, and llama-quantize shrinks a high-precision GGUF to lower-bit formats, which can cost some accuracy.
MLX LM ↗ ★ 7,470
Apple's MLX package for running and fine-tuning models on Apple silicon. It can quantize models and upload them to the Hugging Face Hub, and supports LoRA and full fine-tuning.

Merge it with another model

mergekit ↗ ★ 7,415
Merges pretrained models with many merge methods. It works out-of-core, so merges can run on a CPU or with as little as 8 GB of VRAM, and it also handles LoRA extraction and mixture-of-experts merges.

Learn the techniques

LLM Course ↗ ★ 84,684
A free course with roadmaps and Colab notebooks. Its notebooks include one-click tools for merging with MergeKit, quantizing to GGUF, GPTQ, EXL2, AWQ and HQQ, and automatic abliteration with custom datasets.

Compare all tools →

Source provenance

Downloads: 32
Likes: 9
Metadata checked: 2026-10-26 · 04:15:31 UTC
First seen in index: 2026-10-26 · 04:15:31 UTC
Source revision: c707596609f9a45c7fe3babe30e0d238a92b8364

Source-card excerpt

Source model card excerpt: ✨G3 Glitter 27B✨ A creative writing model based on Gemma 3 27B. Columbidae/gemma-3-27b-half, a 50/50 merge of 27B IT and 27B PT, was used

From allura-org/Gemma-3-Glitter-27B. Creator model-card excerpt, not an independent review.

Inspect before you run

Low refusal is not proof of correctness. Outputs can be inaccurate, harmful, or biased. Review the model card, license and files yourself. A license tag is not commercial-use advice. Parameter counts are source metadata; no hardware compatibility is implied.

This directory does not host or redistribute model weights. Refusal profiles are measured by us on a fixed prompt set; all other behavior claims are source-reported.

Collection rules & limitations ↗