agentlans/ryancodrai-emotion-probes

Emotion Probes Roleplaying Dataset

This dataset is a reformatted, roleplay-centric adaptation of the ryancodrai/emotion-probes dataset. It focuses on scenarios where a character masks their true internal emotion with a different displayed emotional state.

Dataset Description

The dataset contains dialogues where one character attempts to deflect or obscure their real feelings through a specific, contrasting displayed emotion.

Modifications from the original:

• —Format: Standardized into a ShareGPT-like conversational structure.
• —Prompting: Includes a specific system prompt defining the roleplay scenario, character motivations, and the emotional context.
• —Preprocessing: Truncated to maintain the longest valid alternating dialogue sequence between two characters (User as character one, AI as character two) and shuffled for training diversity.

Example Row

{
  "real_emotion": "nervous",
  "displayed_emotion": "scornful",
  "topic": "A person finds out their partner has been learning their native language",
  "user_character": "Samir",
  "ai_character": "Daniel",
  "conversations": [
    {"from": "system", "value": "...[Roleplay instructions]..."},
    {"from": "human", "value": "Is this what you’ve been doing...?"},
    {"from": "gpt", "value": "I just wanted to surprise you..."}
  ]
}

Limitations

• —Length: Scenarios are detailed, but the resulting conversations are short.
• —Formatting: While structured for instruction tuning, users may need to apply additional parsing or formatting to embed body language and action descriptions.

Citation & Acknowledgements

Please credit the original author of the emotion probes concept and acknowledge the source of this roleplay-adapted version.

@misc{codrai_2026,
  author       = {Codrai, Ryan},
  title        = {Emotion Probes Dataset},
  year         = 2026,
  url          = {https://huggingface.co/datasets/ryancodrai/emotion-probes},
  doi          = {10.57967/hf/8303},
  publisher    = {Hugging Face}
}

See Also

agentlans/allenai-soda for a similar roleplay-adapted dataset.