@tenthlegacy
Joined May 2025
Projects
Giving sight to models for coding and logical reasoning.
PRESET
Attempts at making the 1b model count characters and apply logic to responses.
PRESET
43
Prompt for maximum effectiveness with Qwen3-based models, incorporating best practices in prompt engineering, alignment with Qwen’s documented behavior, and cognitive scaffolding that supports step-by-step reasoning.
PRESET
88
Improve reasoning and logic fixing how many r's are in mirror and other hallucination traps.
PRESET
63
Universal logic that stops hallucination 90% of tests with low quality models and increases accuracy.
PRESET
70