Upload folder using huggingface_hub

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+gemma-3-27b-it-qat-abliterated.q2_k.gguf filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,33 @@
+---
+license: gemma
+library_name: transformers
+pipeline_tag: image-text-to-text
+base_model: google/gemma-3-27b-it-qat-q4_0-unquantized
+tags:
+- autoquant
+- gguf
+---
+
+# 💎 Gemma 3 27B IT QAT Abliterated
+
+![image/png](https://cdn-uploads.huggingface.co/production/uploads/61b8e2ba285851687028d395/NjwzenHhKsuPRMPYxyN4p.png)
+<center>Gemma 3 QAT Abliterated <a href="https://huggingface.co/mlabonne/gemma-3-1b-it-qat-abliterated">1B</a> • <a href="https://huggingface.co/mlabonne/gemma-3-4b-it-qat-abliterated">4B</a> • <a href="https://huggingface.co/mlabonne/gemma-3-12b-it-qat-abliterated">12B</a> • <a href="https://huggingface.co/mlabonne/gemma-3-27b-it-qat-abliterated">27B</a></center>
+
+This is an uncensored version of [google/gemma-3-27b-it-qat-q4_0-unquantized](https://huggingface.co/google/gemma-3-27b-it-qat-q4_0-unquantized) created with a new abliteration technique.
+See [this article](https://huggingface.co/blog/mlabonne/abliteration) to know more about abliteration.
+
+This is a new, improved version that targets refusals with enhanced accuracy.
+
+I recommend using these generation parameters: `temperature=1.0`, `top_k=64`, `top_p=0.95`.
+
+## ✂️ Abliteration
+
+![image/png](https://cdn-uploads.huggingface.co/production/uploads/61b8e2ba285851687028d395/xzUdjHWYL0p-KyqlIpN4x.png)
+
+The refusal direction is computed by comparing the residual streams between target (harmful) and baseline (harmless) samples. 
+The hidden states of target modules (e.g., o_proj) are orthogonalized to subtract this refusal direction with a given weight factor. 
+These weight factors follow a normal distribution with a certain spread and peak layer. 
+Modules can be iteratively orthogonalized in batches, or the refusal direction can be accumulated to save memory.
+
+Finally, I used a hybrid evaluation with a dedicated test set to calculate the acceptance rate. This uses both a dictionary approach and [NousResearch/Minos-v1](https://huggingface.co/NousResearch/Minos-v1). 
+The goal is to obtain an acceptance rate >90% and still produce coherent outputs.

gemma-3-27b-it-qat-abliterated.q2_k.gguf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:42bb39a39871eac590c357fca7aa51cf2ba64030f381b0c879fcbb9817551076
+size 10503720704