Post
280
Uncensoring Mistral,
give it a try
s3nh/Ministral-3-14B-Instruct-2512-BF16-abliterated
give it a try
s3nh/Ministral-3-14B-Instruct-2512-BF16-abliterated
š¤ Open to Collab
s3nh
s3nh
AI & ML interests
Quantization, LLMs, Deep Learning for good. Follow me if you like my work. Patreon.com/s3nh
Recent Activity
reacted to danielhanchen's post with š„ about 5 hours ago
Introducing Unsloth for AMD š
You can now train & run LLMs on your AMD hardware
ā¢ We collaborated with AMD to enable you to train & run 500+ models on AMD GPUs
ā¢ Works on Windows, WSL, Linux
ā¢ Train Qwen, Gemma on just 3GB VRAM
GitHub: https://github.com/unslothai/unsloth
Blog + Guide: https://unsloth.ai/docs/basics/amd updated a model 7 days ago
s3nh/Bonsai-27B-unpacked-abliterated-uncensored published a model 7 days ago
s3nh/Bonsai-27B-unpacked-abliterated-uncensoredOrganizations
Posts 12
Post
347
Existing methods ā GPTQ, AWQ, llama.cpp's k-quants ā minimize empirical loss heuristically. None of them prove they are optimal in any information-theoretic sense. ICRB-Q builds a quantization scheme that is provably optimal via the CramĆ©r-Rao lower bound (CRB): no unbiased estimator of a weight can have lower variance than [F(Īø)]ā»Ā¹, where F is the Fisher information matrix.
GGUF converted models, sorted by creation date.
-
s3nh/AdaptLLM-law-LLM-13B-GGUF
Text Generation ā¢ 13B ā¢ Updated ā¢ 55 ā¢ 2 -
s3nh/AdaptLLM-finance-LLM-13B-GGUF
Text Generation ā¢ 13B ā¢ Updated ā¢ 44 ā¢ 6 -
s3nh/AdaptLLM-medicine-LLM-13B-GGUF
Text Generation ā¢ 13B ā¢ Updated ā¢ 47 ā¢ 1 -
s3nh/decapoda-research-Antares-11b-v1-GGUF
Text Generation ā¢ 11B ā¢ Updated ā¢ 94 ā¢ 1
GGUF converted models, sorted by creation date.
-
s3nh/AdaptLLM-law-LLM-13B-GGUF
Text Generation ā¢ 13B ā¢ Updated ā¢ 55 ā¢ 2 -
s3nh/AdaptLLM-finance-LLM-13B-GGUF
Text Generation ā¢ 13B ā¢ Updated ā¢ 44 ā¢ 6 -
s3nh/AdaptLLM-medicine-LLM-13B-GGUF
Text Generation ā¢ 13B ā¢ Updated ā¢ 47 ā¢ 1 -
s3nh/decapoda-research-Antares-11b-v1-GGUF
Text Generation ā¢ 11B ā¢ Updated ā¢ 94 ā¢ 1
Collection of LLMs finetuned using axolotl library, mostly
models 273
s3nh/Bonsai-27B-unpacked-abliterated-uncensored
27B ā¢ Updated ā¢ 71 ā¢ 3
s3nh/Tess-4-27B-abliterated
Image-Text-to-Text ā¢ 27B ā¢ Updated ā¢ 116 ā¢ 2
s3nh/Ministral-3-14B-Instruct-2512-BF16-abliterated
14B ā¢ Updated ā¢ 108
s3nh/Ministral-3-8B-Instruct-2512-BF16-abliterated
9B ā¢ Updated ā¢ 109
s3nh/fable-traces-abliterated
Text Generation ā¢ 4B ā¢ Updated ā¢ 446
s3nh/FastContext-1.0-4B-SFT-abliterated
Text Generation ā¢ 4B ā¢ Updated ā¢ 27 ā¢ 1
s3nh/LFM2.5-350M-abliterated
Text Generation ā¢ 0.4B ā¢ Updated ā¢ 28 ā¢ 2
s3nh/OProver-8B-Base-GGUF
8B ā¢ Updated ā¢ 16
s3nh/BitCPM4-CANN-3B-GGUF
4B ā¢ Updated ā¢ 129 ā¢ 1
s3nh/Nanbeige-4.1-3B-Uncensored
Text Generation ā¢ 4B ā¢ Updated ā¢ 11 ā¢ 1