πŸ’  SteveAI-Mistral-Personas (1,000 Souls)

This is a specialized Mistral-7B-v0.3 model featuring the unique SteveAI Qubit-State Architecture. It contains 1,000 distinct personality vectors ("souls") that can be swapped in real-time by updating specific parameters in the MLP layers.

πŸš€ Technical Highlights

  • Architecture: SteveAI_SyncMLP custom layer injection.
  • Precision: Achieved a training loss of 0.00000383.
  • Training: Orchestrated on dual NVIDIA T4 GPUs using a custom synchronization protocol to prevent device-mismatch errors.
  • Identity Swapping: Near-instantaneous (copying a single vector into the MLP state) rather than full weight reloading.

🎭 Featured Personas

Name Background Perspective
Xyla-Xen Conceptual Artist Views the abyss as an abstract sky of blue light and bioluminescence.
Serafina Mote 14th Century Diver A historical "time-traveler" logic blending alchemy with modern rebreathers.
Lady Evangeline Aristocratic Explorer Uses dense, Victorian-era scientific terminology to describe hadal zones.
Jax Kincaid Deep-Sea Biologist Focused on biochemistry and crustacean symbiosis.

πŸ§ͺ Recommended Settings

To prevent the "Looping" or "Collapse" behavior sometimes found in high-precision states:

  • Repetition Penalty: 1.2
  • No Repeat N-Gram Size: 3
  • Temperature: 0.7 - 0.85
  • Top_P: 0.9

πŸ’» About SteveAI

Part of the SteveAI ecosystem, hosted at steve-ai.netlify.app. Developed by Saadpie (Saad Abdulrehman).

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. πŸ™‹ Ask for provider support

Model tree for saadpie/SteveAI-Mistral-Personas

Finetuned
(381)
this model