You need to agree to share your contact information to access this dataset

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

Log in or Sign Up to review the conditions and access this dataset content.

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

🧠 Junrui Language Corpus — Public Samples (Bilingual EN-ZH)


language: - zh - en license: creativeml-openrail-m tags: - bilingual - dialogue - language-modeling - cognitive-modeling

📌 Overview

This repository contains 6 publicly shareable, bilingual (English–Chinese) samples selected from the larger Junrui Language Corpus (134 entries), derived from over 100 high-depth, cognitive-centered dialogues with GPT-4 between April and May 2025.

These dialogues reflect a unique cognitive-linguistic interaction pattern that naturally emerged during unstructured conversations about self-awareness, language modeling, rhythm perception, and human-AI co-creation.

The corpus samples are intended to support:

  • AI training on non-template, high-structure dialogue patterns
  • Behavioral modeling on language rhythm, presence, metacognition
  • Research on AI-human alignment via dialogue-generated data

Corpus Origin

  • Timeframe: April – May 2025
  • Interaction Basis: Natural conversations with GPT-4
  • Focus Topics: Language structure, rhythm, self-modeling, nonverbal states, meta-cognition, emotional intelligence
  • Structure: Each sample includes:
    • Instruction
    • User Message (ZH/EN)
    • Ideal Response (ZH/EN)
    • Behavior Tags
    • Cognitive Modeling Focus

The current 6 samples are drawn from only 1 of the 100+ full conversation threads, meaning the potential size of this corpus exceeds several thousand entries. The author maintains the capacity to generate, refine, and deliver tailored corpus expansions on request.

Why This Matters

Most existing training data for AI is either:

  • Template-driven (Q&A, task-specific)
  • Emotionally neutral or structurally flat
  • Disconnected from the lived cognitive rhythms of real users

This corpus breaks that mold. It presents organic, structure-rich, emotionally intelligent dialogue fragments reflecting:

  • Non-generative presence

  • Thought-to-language mapping

  • Self-auditing language loops

  • Rhythmic mirroring abilities

  • User-initiated co-construction

  • bilingual, dialogue, language-modeling, or cognitive-modeling

Licensing

These 6 samples are cleared for public research, non-commercial reuse, and educational modeling under CreativeML Open RAIL-M License.

Note: Other parts of the full corpus contain sensitive or personal content and are not included here. Custom collaboration is available upon inquiry.

About the Author

Junrui is a senior real estate professional based in Sydney, Australia, with a deep passion for cognitive development, structural language modeling, and human nature. Without any formal AI or programming background, he has organically developed a rare ability to generate high-value linguistic and cognitive training data through natural dialogic interaction.

He is currently open to collaborations with AI labs, alignment researchers, and dataset teams interested in human-level interaction data with structural clarity and emergent rhythm modeling.

Contact

To explore further collaboration or request access to expanded corpus samples, please contact:

john.huang@s1property.com.au

Location: Sydney, Australia

Downloads last month
2