A collection of models trained on Opus distillation data, in efforts to transfer better reasoning.
Convergent Intelligence PRO
reaperdoesntknow
AI & ML interests
About Us
Mission
Convergent Intelligence advances original research in discrepancy calculus, adaptive systems, and applied AI, translating those insights into client controls, playbooks, and leadership-ready briefs.
Recent Activity
updated a model about 6 hours ago
reaperdoesntknow/reaperdoesntknow updated a model about 6 hours ago
reaperdoesntknow/TopologicalQwen updated a model about 6 hours ago
reaperdoesntknow/TameForCasualLMOrganizations
The first defense AI reasoning model's on Hugging Face.
Proof-weighted distillation, Qwen3-30B to 1.7B/0.6B. Three teachers: Instruct, Thinking, Coder. The core method series. DOI 10.57967/hf/8165
-
reaperdoesntknow/Qwen3-1.7B-Distilled-30B-A3B
Text Generation • 2B • Updated • 2.45k • 2 -
reaperdoesntknow/Qwen3-1.7B-Distilled-30B-A3B-SFT-GGUF
Text Generation • 2B • Updated • 3.12k -
reaperdoesntknow/Qwen3-1.7B-Distilled-30B-A3B-SFT
2B • Updated • 257 -
reaperdoesntknow/Qwen3-0.6B-Distilled-30B-A3B
Text Generation • 0.8B • Updated • 2.33k
SAGI is a novel causal language model that integrates **swarm intelligence dynamics** with transformer architecture.
Symbolic-transformer hybrids with persistent memory, 1B to 14B. Experimental research checkpoints.
Various models trained with our heterogeneous optimizer designed for efficient parameter routing across large projection matrices.
One model, two internal voices: explore, examine, respond. Self-critique reasoning on shared weights. DOI 10.57967/hf/8184
Qemma is a HuggingFace-native hybrid model that merges Gemma-3 (1B) and Qwen-3 (0.6B) and (1.7B) using Gap Envelope Integrals .
A collection of models trained on Opus distillation data, in efforts to transfer better reasoning.
Various models trained with our heterogeneous optimizer designed for efficient parameter routing across large projection matrices.
The first defense AI reasoning model's on Hugging Face.
One model, two internal voices: explore, examine, respond. Self-critique reasoning on shared weights. DOI 10.57967/hf/8184
Proof-weighted distillation, Qwen3-30B to 1.7B/0.6B. Three teachers: Instruct, Thinking, Coder. The core method series. DOI 10.57967/hf/8165
-
reaperdoesntknow/Qwen3-1.7B-Distilled-30B-A3B
Text Generation • 2B • Updated • 2.45k • 2 -
reaperdoesntknow/Qwen3-1.7B-Distilled-30B-A3B-SFT-GGUF
Text Generation • 2B • Updated • 3.12k -
reaperdoesntknow/Qwen3-1.7B-Distilled-30B-A3B-SFT
2B • Updated • 257 -
reaperdoesntknow/Qwen3-0.6B-Distilled-30B-A3B
Text Generation • 0.8B • Updated • 2.33k
SAGI is a novel causal language model that integrates **swarm intelligence dynamics** with transformer architecture.
Qemma is a HuggingFace-native hybrid model that merges Gemma-3 (1B) and Qwen-3 (0.6B) and (1.7B) using Gap Envelope Integrals .
Symbolic-transformer hybrids with persistent memory, 1B to 14B. Experimental research checkpoints.