EldritchLabs/MN-12B-Mag-Mell-R1-Uncensored-Scale1.2

Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:12BQuant:FP8Ctx Length:32kPublished:Mar 5, 2026Architecture:Transformer0.0K Warm

MN-12B-Mag-Mell-R1-Uncensored-Scale1.2 is a 12 billion parameter language model developed by EldritchLabs, based on the Naphula/MN-12B-Mag-Mell-R1-Uncensored model. It features a context length of 32768 tokens and is specifically optimized with a scale of 1.2, which is noted as a sweet spot for Nemo models. This uncensored model is designed for applications requiring unrestricted content generation and potentially enhanced intelligence compared to its predecessor.

Loading preview...

MN-12B-Mag-Mell-R1-Uncensored-Scale1.2 Overview

EldritchLabs' MN-12B-Mag-Mell-R1-Uncensored-Scale1.2 is a 12 billion parameter language model derived from the Naphula/MN-12B-Mag-Mell-R1-Uncensored base. This iteration maintains the uncensored nature of its predecessor while introducing a key modification: a scale factor of 1.2. This specific scaling is highlighted as an optimal configuration for Nemo-based models, aiming to deliver improved performance.

Key Capabilities

  • Unrestricted Content Generation: Designed to produce uncensored outputs, suitable for use cases requiring broad content flexibility.
  • Optimized Performance: Utilizes a 1.2 scale factor, which is identified as a "sweet spot" for Nemo models, potentially leading to enhanced intelligence and efficiency.
  • Large Context Window: Supports a context length of 32768 tokens, enabling processing of extensive inputs and generating coherent long-form content.

Good For

  • Applications requiring a highly flexible and uncensored language model.
  • Scenarios where the 1.2 scale factor might offer a performance advantage over other scaling configurations.
  • Tasks benefiting from a substantial context window for complex or lengthy interactions.

Popular Sampler Settings

Top 3 parameter combinations used by Featherless users for this model. Click a tab to see each config.

temperature
top_p
top_k
frequency_penalty
presence_penalty
repetition_penalty
min_p