MuXodious/WeirdCompound-v1.7-24b-absolute-heresy-noslop-v2

TEXT GENERATIONPricing:Input $0.7 / Cached $0.04 / Output $1.16Concurrent Unit Cost:2Model Size:24BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Feb 15, 2026Architecture:Transformer0.0K Featherless Exclusive Cold

The MuXodious/WeirdCompound-v1.7-24b-absolute-heresy-noslop-v2 is a 24 billion parameter language model with a 32768 token context length, created by MuXodious. This model is a multi-stage merge of several fine-tuned models, including Cydonia, Eurydice, and Dans-PersonalityEngine, which has undergone a 'noslopfication' process using P-E-W's Heretic engine to reduce 'slop' and 'refusals'. It is specifically engineered to exhibit an 'Absolute Heresy' classification, indicating a significant reduction in model refusals and a low KL Divergence, making it distinct for its altered behavioral characteristics.

Loading preview...

Model Overview

MuXodious/WeirdCompound-v1.7-24b-absolute-heresy-noslop-v2 is a 24 billion parameter language model with a 32768 token context length, developed by MuXodious. This model is a multi-stage merge of various specialized fine-tunes, including TheDrummer/Cydonia-24B-v4.2.0, aixonlab/Eurydice-24b-v3.5, and PocketDoc/Dans-PersonalityEngine-V1.3.0-24b, among others. The merging process utilized methods like Model Stock, SLERP, and NuSLERP.

Unique Characteristics

This model has undergone a unique 'noslopfication' and 'heretication' process using P-E-W's Heretic engine. This process aims to significantly reduce 'slop' and 'refusals' in the model's output. Key metrics from this process include:

  • Slop Score: 12/100 (down from an initial 90/100)
  • Refusals: 5/100 (down from an initial 99/100)
  • KL Divergence: 0.0446 (for noslopfication) and 0.0330 (for heretication)

These results classify the model as 'Absolute Heresy', indicating a deliberate alteration of its default refusal behavior and a low divergence from its base. The developer notes that the model "might be lobotomised" due to these processes, suggesting a highly specialized and potentially unconventional output style.

Potential Use Cases

Given its unique 'heretication' and 'noslopfication' for reduced refusals, this model could be suitable for experimental applications where a highly unconstrained or unconventional response style is desired. Developers seeking to explore the boundaries of LLM behavior, particularly in creative or non-standard content generation, might find this model useful. It is important to test its specific outputs for suitability in any given application due to its intentionally altered characteristics.

Finetunes
1 models