OccultAI/Adversary-8B-v1a

TEXT GENERATIONConcurrency Cost:1Model Size:8BQuant:FP8Ctx Length:32kTool Calling:SupportedPublished:May 9, 2026Architecture:Transformer0.0K Cold

Adversary-8B-v1a is an 8 billion parameter LlamaForCausalLM architecture model developed by OccultAI, created through a 'della' merge of several 8B base models including Llama-3.1-Nemotron-8B-UltraLong-1M-Instruct and others. This model is designed to combine diverse characteristics from its constituent models, aiming for a broad range of capabilities. It leverages a 32768 token context length, making it suitable for tasks requiring extensive contextual understanding.

Loading preview...

Adversary-8B-v1a: A Merged 8B Language Model

Adversary-8B-v1a is an 8 billion parameter language model developed by OccultAI, built upon the LlamaForCausalLM architecture. This model is a product of a sophisticated 'della' merge operation, combining several distinct 8B base models to synthesize their individual strengths.

Key Characteristics

  • Architecture: Based on the LlamaForCausalLM framework.
  • Merging Strategy: Utilizes the 'della' merge method, which combines multiple base models with specific weighting parameters for each component.
  • Constituent Models: The merge incorporates diverse models such as:
    • SicariusSicariiStuff--Llama-3.1-Nemotron-8B-UltraLong-1M-Instruct_Abliterated
    • DarkArtsForge--Raven-8B-v1 (associated with Edgar Allan Poe)
    • OccultAI--Barbot-8B-v1 (associated with Douglas Adams)
    • EldritchLabs--Cthulhu-8B-v1.4 (associated with H.P. Lovecraft)
    • OccultAI--Morpheus-8B-v1
    • SicariusSicariiStuff--Assistant_Pepe_8B
  • Context Length: Supports a substantial context window of 32768 tokens.
  • Tokenizer: Employs the tokenizer from OccultAI--Morpheus-8B-v1.

Potential Use Cases

Given its merged nature, Adversary-8B-v1a is likely intended for applications that benefit from a blend of different model characteristics, potentially excelling in tasks requiring:

  • Diverse stylistic generation: Drawing from models associated with distinct literary figures.
  • Extended context understanding: Due to its 32768 token context length.
  • General-purpose language tasks: Leveraging the broad capabilities of its Llama-based foundation and merged components.

Finetunes
1 models