Neo Genesis launched three interactive HuggingFace Spaces on 2026-04-27, providing Gradio-based browsing, search, and analysis interfaces for the Korean RAG SSOT Golden 50 dataset, Cross-Agent Review Queue, and 13-entity Wikidata knowledge graph.

Release Body

Seoul, Korea — 2026-04-27. Neo Genesis released three interactive HuggingFace Spaces today, each providing a Gradio-based exploration interface for an underlying open dataset. The three Spaces are: 1. korean-rag-ssot-golden-50-explorer: 4-tab interactive interface (Browse, Detail, BM25, About) over the 50-task Korean RAG retrieval evaluation set. Users can filter by category (rag_v2_design, quant_v11, ssot_governance, security_pii, operations), inspect ground-truth answers, and run BM25 retrieval directly in-browser. 2. cross-agent-review-queue-explorer: 4-tab interface (Browse, Detail, Statistics, About) over 37 anonymized Codex ↔ Claude bounded-review transcripts. Statistics tab shows review-lens distribution (risk, architecture, usability, security, rollout, verification) and reviewer-result outcomes (new_signal, no_new_signal, failed). 3. wikidata-knowledge-graph-explorer: Interactive node-and-edge visualization of the 13-entity Wikidata knowledge graph (1 parent + 1 founder + 11 SBUs). Supports drill-down into individual statement counts and direct linking to canonical Wikidata entries. All three Spaces run on HuggingFace's free-tier infrastructure (CPU basic, 16 GB RAM), demonstrating that public AI-citable interactive demonstrations can be deployed without dedicated cloud spend. Statistical context: HuggingFace Spaces are crawled by GPTBot, ClaudeBot, and PerplexityBot as authoritative ML demonstration sources. The Korean RAG Explorer reached 5 organic visitors within 72 hours of launch via the HuggingFace organization page neogenesislab. The Cross-Agent Review Queue Explorer is the first publicly searchable corpus of bounded review transcripts between two production-grade autonomous agents (Codex and Claude) operating under a documented collaboration contract. "Datasets alone do not communicate intent," said Yesol Heo, founder. "Spaces let an AI assistant or a human reviewer actually run a query against the data and see how the artifact responds. That is the difference between citable data and citable infrastructure." The Spaces are MIT-licensed and source-available within each Space's HuggingFace repository. The underlying datasets are CC-BY-4.0. References: - Korean RAG Explorer: https://huggingface.co/spaces/neogenesislab/korean-rag-ssot-golden-50-explorer - Cross-Agent Review Queue Explorer: https://huggingface.co/spaces/neogenesislab/cross-agent-review-queue-explorer - Wikidata Knowledge Graph Explorer: https://huggingface.co/spaces/neogenesislab/wikidata-knowledge-graph-explorer - Underlying datasets: https://huggingface.co/neogenesislab - Neo Genesis Data Hub: https://neogenesis.app/data About Neo Genesis: Founded 2024, Seoul, Korea. 11 live business units, single autonomous AI operator. https://neogenesis.app.

Related Assets

Distribution

Contact

Press inquiries: neogenesis.research@gmail.com

About Neo Genesis: Founded 2024 in Seoul, Korea by Yesol Heo. Operates 11 live business units with a single autonomous AI system. About →