reaperdoesntknow/Symbiotic-1B

Hugging Face
TEXT GENERATIONConcurrent Unit Cost:1Model Size:0.8BQuant:BF16Context Size:32kTool Calling:SupportedPublished:May 6, 2025License:afl-3.0Architecture:Transformer0.0K Featherless Exclusive Warm

SymbioticLM-1B by reaperdoesntknow is a 1 billion parameter hybrid symbolic-transformer model, built on a Qwen-1B backbone. It integrates a symbolic processing pipeline and persistent episodic memory for enhanced reasoning. This model is optimized for lightweight, memory-augmented symbolic inference on CPU and embedded systems, excelling in areas like educational agents, graph-based explanations, and procedural planning.

Loading preview...

SymbioticLM-1B: Hybrid Symbolic-Transformer for Lightweight Reasoning

SymbioticLM-1B, developed by reaperdoesntknow, is a compact 1 billion parameter model that uniquely combines a Qwen-1B rotary transformer backbone with a sophisticated symbolic processing pipeline and persistent episodic memory. This architecture is designed to provide advanced reasoning capabilities in resource-constrained environments, making it suitable for CPU and embedded inference.

Key Capabilities & Architecture

  • Hybrid Design: Fuses a Qwen-1B transformer with a symbolic cognitive engine, including symbolic memory, dynamic thought evolution, and entropy-gated control.
  • Symbolic Processing: Features specialized symbolic modules like ThoughtDynamicsLNN, CrystallineProcessor (DNAConv GNN), LiquidThoughtProcessor, and HelicalDNAProcessor.
  • Memory Augmentation: Incorporates 2048 symbolic vectors with entropic and contextual retrieval, enhancing its reasoning over time.
  • Dream Mode: Includes a "Dream Mode" for symbolic simulation via a ThoughtGenerator.
  • Discrepancy Calculus Foundation: Developed under the Discrepancy Calculus (DISC) framework, which treats training singularities as structural signals for understanding learning problems.

Intended Uses

  • CPU-optimized symbolic inference.
  • Educational agents requiring memory and reasoning.
  • Generation of graph-based explanations.
  • Procedural planning, mathematical modeling, and small-scale code generation.

Limitations

  • Less fluent in free-form language compared to larger, purely generative models.
  • Symbolic accuracy benefits from careful memory curation.
  • Complex queries in "Dream Mode" may require warm-up or symbolic seeding.

Finetunes
1 models