CITATION.cff

cff-version: 1.2.0
title: "CHEN: Collaborative Heterogeneous Expert Network"
message: >-
  If you use CHEN in your research, please cite it using the metadata
  below. The project is released into the public domain under CC0 1.0.
type: software
authors:
  - given-names: "CHEN"
    family-names: "Contributors"
    affiliation: "Community"
repository-code: "https://github.com/your-org/chen"
url: "https://github.com/your-org/chen"
abstract: >-
  CHEN is a distributed inference architecture that replaces a single
  monolithic hyper-scale model with a coordinated network of specialized,
  low-parameter models (3B-8B). It routes tokens through a dynamic pipeline
  and shares latent memory states (KV-caches) between models, simulating
  the parameter capacity and reasoning depth of a frontier model at a
  fraction of the compute cost. The repository implements three pillars —
  the Latent State Router, Specialized Micro-Experts, and Shared External
  Memory — across three experimental phases (static cascade, KV-cache
  passing, and dynamic routing), with a benchmark harness measuring
  Effective Parameter Utilization (EPU), cost-per-1M-tokens, and
  latency-to-accuracy ratio.
keywords:
  - distributed-inference
  - mixture-of-experts
  - kv-cache
  - small-language-models
  - routing
  - rag
  - llm
  - decentralized-ai
license: CC0-1.0
version: "0.1.0"
date-released: "2025-01-15"
preferred-citation:
  type: software
  title: "CHEN: Collaborative Heterogeneous Expert Network"
  authors:
    - given-names: "CHEN"
      family-names: "Contributors"
  year: 2025
  url: "https://github.com/your-org/chen"