LLM EXPLORER 59,420 MODELS INDEXED

Sensualize Mixtral GPTQ by TheBloke

By TheBloke · 9 downloads
HF Repo ↗ Copy URL

Sensualize Mixtral GPTQ is an open-source language model by TheBloke. Features: 46.7b LLM, VRAM: 23.8GB, Context: 32K, License: cc-by-nc-4.0, Quantized, LLM Explorer Score: 0.1.

  4-bit Base model:quantized:sao10k/se... Base model:sao10k/sensualize-m... Dataset:nobodyexistsontheinter...   Gptq   Mixtral   Quantized   Region:us   Safetensors
Model Card on HF 🤗: https://huggingface.co/TheBloke/Sensualize-Mixtral-GPTQ 

Sensualize Mixtral GPTQ Parameters and Internals

Model Type 
mixtral
Use Cases 
Areas:
Roleplay, ERP (Erotic Roleplay)
Considerations:
Trained on NSFW data for specific use cases. Considerations for roleplay should be taken into account.
Additional Notes 
An experimental model with finicky performance, performs well with right prompts and settings. Quantized versions available for VRAM optimization.
Training Details 
Data Sources:
NobodyExistsOnTheInternet/full120k, NSFW Instruct & De-Alignment Data
Data Volume:
80M tokens over 1 epoch
Methodology:
Trained using Alpaca format, megablocks-based fork of transformers, and Charles Goddard's ZLoss.
Training Time:
12 hours
Hardware Used:
2xA100 GPUs at batch size 5, grad 5
Model Architecture:
Based on mistralai/Mixtral-8x7B-v0.1
Input Output 
Input Format:
Instruction-Input-Response
Accepted Modalities:
text
Performance Tips:
With the right settings, the model can perform well. Recommended settings include Universal-Light or Universal-Creative in SillyTavern.
LLM NameSensualize Mixtral GPTQ
Repository 🤗https://huggingface.co/TheBloke/Sensualize-Mixtral-GPTQ 
Model NameSensualize Mixtral
Model CreatorSaofiq
Base Model(s)  Sensualize Mixtral Bf16   Sao10K/Sensualize-Mixtral-bf16
Model Size46.7b
Required VRAM23.8 GB
Updated2026-07-11
MaintainerTheBloke
Model Typemixtral
Model Files  23.8 GB
GPTQ QuantizationYes
Quantization Typegptq
Model ArchitectureMixtralForCausalLM
Licensecc-by-nc-4.0
Context Length32768
Model Max Length32768
Transformers Version4.37.0.dev0
Tokenizer ClassLlamaTokenizer
Padding Token</s>
Vocabulary Size32000
Torch Data Typebfloat16

Best Alternatives to Sensualize Mixtral GPTQ

Best Alternatives
Context / RAM
Downloads
Likes
Mixtral 8x7B V0.1 GPTQ32K / 23.8 GB181126
...ixtral 8x7B Instruct V0.1 GPTQ32K / 23.8 GB210141
...Hermes 2 Mixtral 8x7B DPO GPTQ32K / 23.8 GB3126
...Hermes 2 Mixtral 8x7B SFT GPTQ32K / 23.8 GB1011
Dolphin 2.5 Mixtral 8x7b GPTQ32K / 23.8 GB81114
Bagel DPO 8x7b V0.2 GPTQ32K / 23.8 GB122
Open Gpt4 8x7B V0.2 GPTQ32K / 23.8 GB96
...xtral Instruct 8x7b Zloss GPTQ32K / 23.8 GB72
....1 LimaRP ZLoss DARE TIES GPTQ32K / 23.8 GB56
Dolphin 2.7 Mixtral 8x7b GPTQ32K / 23.8 GB3219
Note: green Score (e.g. "73.2") means that the model is better than TheBloke/Sensualize-Mixtral-GPTQ.