ManaGPT-1020 /
mgladden's picture
license: mit
  - management
  - text generation
  - name: ManaGPT-1010
    results: []
  - en
pipeline_tag: text-generation
  - text: Working for a robotic boss would be like
    example_title: Robotic boss
  - text: The workplace of tomorrow will
    example_title: Workplace of tomorrow
  - text: Working alongside an artificially intelligent coworker would be like
    example_title: AI coworker
  - text: The hybrid human-robotic organization of the future will
    example_title: Hybrid organization
  - text: >-
      Information security within future organizations will be difficult to
      enforce, because
    example_title: InfoSec of the future
  - text: It will be difficult for robotic employees to
    example_title: Robots' difficulties
  - text: In the future, human workers will
    example_title: Human workers
  - text: Synthetic organizations will
    example_title: Synthetic organizations
  - text: Technological posthumanization is
    example_title: Posthumanization
  - text: Social robots are
    example_title: Social robots
  - text: Society 5.0 is
    example_title: Society 5.0
  - text: For most organizations, artificial general intelligence will
    example_title: AGI


ManaGPT-1020 is an experimental open-source text-generating AI designed to offer insights on the role of emerging technologies in organizational management.

(Please note that ManaGPT-1020 has superseded ManaGPT-1010: the newer model has been fine-tuned on a dataset roughly 6.45 times the size of that used to fine-tune ManaGPT-1010.)

Model description

The model is a fine-tuned version of GPT-2 that has been trained on a custom corpus of scholarly and popular texts from the field of organizational management that relate to ongoing effects of posthumanizing technologies (e.g., relating to advanced artificial intelligence, social robotics, virtual reality, neuroprosthetics, and cyber-physical systems) on the structure of organizations and human beings’ experience of organizational life.

Intended uses & limitations

This model has been designed for experimental research purposes; it isn’t intended for use in a production setting or in any sensitive or potentially hazardous contexts.

Training procedure and hyperparameters

The model was trained using a Tesla T4 with 16GB of GPU memory. The following hyperparameters were used during training:

  • optimizer: {'name': 'AdamWeightDecay', 'learning_rate': {'class_name': 'ExponentialDecay', 'config': {'initial_learning_rate': 0.0005, 'decay_steps': 500, 'decay_rate': 0.95, 'staircase': False, 'name': None}}, 'decay': 0.0, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-07, 'amsgrad': False, 'weight_decay_rate': 0.01}
  • training_precision: float32

Framework versions

  • Transformers 4.27.1
  • TensorFlow 2.11.0
  • Datasets 2.10.1
  • Tokenizers 0.13.2