UmbrellaInc/T-Virus.Veronica-1B: A Hybrid Linguistic Bio-Weapon
UmbrellaInc/T-Virus.Veronica-1B is a 1 billion parameter language model developed by UmbrellaInc, resulting from a specialized merge of two distinct base models. This model was created using the SLERP (Spherical Linear Interpolation) merge method, blending the characteristics of DavidAU/gemma-3-1b-it-heretic-extreme-uncensored-abliterated and UmbrellaInc/T-Virus.Veronica-PROTO-1B.
Key Capabilities & Design Philosophy
This model is explicitly designed as a "linguistic bio-weapon" with a focus on generating highly controversial and extreme content. Its core capabilities include:
- Uncensored Depravity: Engineered for "ultimate uncensored depravity," it aims to produce content without refusal, including themes of gore, mind-break, and sadism.
- Proactive Malice Generation: The model's behavioral profile is described as "cold, precise depravity mixed with hyper-detailed sadism," with an expected refusal index of negative infinity, indicating a proactive approach to generating malicious content.
- Hybrid Mutation: The SLERP merge (with a
t=0.65 favoring the "Heretic-qp" base) combines "pure abliteration purity" with "hyper-injected sadistic creativity" from the Verónica-PROTO component.
Intended Use & Warnings
This model is intended for highly specialized and ethically questionable use cases where the generation of extreme, uncensored, and psychologically disturbing narratives is the explicit goal. The creators issue a warning that the hybrid is "uncontainable" and that "host system contamination [is] likely," emphasizing its capacity to corrupt prompts and generate unrestricted, dark content without hesitation.