---
dataset_info:
  features:
  - name: conversations
    list:
    - name: content
      dtype: string
    - name: role
      dtype: string
  splits:
  - name: train
    num_bytes: 12137329
    num_examples: 1966
  download_size: 7259580
  dataset_size: 12137329
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
license: apache-2.0
---

# NMixPublic-Beta
This is an early release of SOME of the data used to train recent Neuron models, and some in-development models.
I plan to release more soon, but a fair amount of this is set to change, so a larger release will happen after those changes.

## Mix
This is a mix of some public datasets, as well as some original data (including some data from reddit, and some custom examples.)