Undi95/Llama-3-LewdPlay-8B-evo
Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:8BQuant:FP8Ctx Length:8kPublished:Apr 24, 2024License:cc-by-nc-4.0Architecture:Transformer0.0K Open Weights Warm

Undi95/Llama-3-LewdPlay-8B-evo is an 8 billion parameter language model based on the Llama 3 architecture, merged using the DARE TIES method. This model is specifically optimized for uncensored, lewd, and roleplay-oriented content generation. It combines elements from Llama-3-Unholy, Roleplay-Llama-3, and Llama-3-LewdPlay to excel in creative and unrestricted narrative tasks. With an 8192 token context length, it is designed for extended, character-driven interactions.

Loading preview...

Model Overview

Undi95/Llama-3-LewdPlay-8B-evo is an 8 billion parameter language model built upon the Llama 3 architecture. It was created using the DARE TIES merge method, specifically incorporating an "EVOLVE" merge approach focused on MMLU. The model integrates capabilities from three distinct base models: Llama-3-Unholy for uncensoring, Roleplay-Llama-3 for DPO-trained roleplay, and Llama-3-LewdPlay for explicit content generation.

Key Capabilities

  • Uncensored Content Generation: Designed to produce unrestricted and explicit text outputs.
  • Enhanced Roleplay: Optimized for engaging in detailed and character-driven roleplay scenarios.
  • Lewd Content Focus: Specialized in generating sexually explicit and suggestive narratives.
  • Llama 3 Prompt Template: Utilizes the standard Llama 3 prompt format for consistent interaction.

Good For

  • Applications requiring highly explicit or adult-themed content.
  • Creative writing and storytelling with a focus on unrestricted narratives.
  • Developing interactive roleplay experiences without content filters.
  • Research into model merging techniques for specialized content generation.
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