| # psy_conv_gpt2 | |
| This dataset is part of the `psy_conv` series, derived from merging multiple psychology-related datasets. It is structured to provide conversational context in mental health, behavioral science, and psychology. | |
| ## Dataset Details | |
| - **Source**: Merged from `Gragroo/psy_gpt_train_validV1_no_names`, `Gragroo/psychology-question-answer_psygpt_with_validation`, and `Gragroo/psychology-10k_psygpt_with_validation` | |
| - **Splits**: Train (80%) and Validation (20%) | |
| - **Fields**: Only contains the `conversations` field in the following format: | |
| ```json | |
| [ | |
| { "from": "human", "value": "<user's input>" }, | |
| { "from": "assistant", "value": "<assistant's response>" } | |
| ] | |
| ``` | |
| ## Usage | |
| This dataset is suitable for training and fine-tuning conversational AI models on psychology and mental health topics. Each conversation is anonymized and structured to provide clear interaction flows between human and assistant. | |
| ## License | |
| This dataset is made available under the license of the original datasets. | |