KaraKaraWitch/Llama-3.3-CURSEDMAGICALGIRL-2.5

Hugging Face
TEXT GENERATIONPricing:Input $2.6 / Cached $0.52 / Output $3Concurrent Unit Cost:4Model Size:70BQuant:FP8Context Size:32kTool Calling:SupportedArchitecture:Transformer0.0K Featherless Exclusive Warm

KaraKaraWitch/Llama-3.3-CURSEDMAGICALGIRL-2.5 is a 70 billion parameter language model with a 32768 token context length. This model is presented with a corrupted system interface, suggesting an unconventional or experimental nature. Its primary differentiator lies in its unique, 'cursed' operational state, implying a specialized or niche application rather than general-purpose use. Developers should consider this model for use cases requiring interaction with or simulation of corrupted or unconventional AI systems.

Loading preview...

Overview

KaraKaraWitch/Llama-3.3-CURSEDMAGICALGIRL-2.5 is a 70 billion parameter language model with an extended context length of 32768 tokens. Unlike typical models, its presentation is characterized by a 'SYSTEM CORRUPTION DETECTED' message and a 'CURSED-MAGICALGIRLS-2.5' error code, indicating a deliberate, unconventional design.

Key Characteristics

  • Unconventional Interface: The model's README explicitly states a corrupted system state, suggesting it operates outside standard LLM paradigms.
  • Specialized Nature: The 'CURSED-MAGICALGIRLS-2.5' designation points to a highly specific, potentially experimental or themed application.
  • Limited Documentation: The model creator has intentionally withheld a standard model card description, reinforcing its unique and non-traditional status.

Potential Use Cases

  • Simulating Corrupted Systems: Ideal for projects requiring an AI that behaves as if it's experiencing system errors or unusual states.
  • Narrative Generation: Could be used in creative writing or game development to generate text from a 'cursed' or glitching AI perspective.
  • Experimental AI Research: Suitable for researchers exploring the boundaries of LLM behavior and interaction with intentionally introduced anomalies.

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