Synthetic Japanese Roleplay Chat Messages in OpenAI Format

A dataset of 5,000 synthetic Japanese roleplay conversations formatted for OpenAI's messages API. The data was generated using the OpenRouter gpt-5-chat model by Aratako and last updated on August 19, 2025. It is a reformatted version of the original dataset 'Synthetic-Japanese-Roleplay-NSFW-gpt-5-chat-5k'.

Use Cases

• Fine-tuning language models for Japanese conversational agents based on synthetic roleplay dialogues.
• Benchmarking model performance on NSFW (Not Safe For Work) content handling in Japanese.
• Studying the structure and patterns of AI-generated roleplay conversations.
• Developing chatbots for interactive entertainment or narrative applications.

Strengths

• Contains 5,000 formatted conversation examples.
• Data is structured for direct compatibility with OpenAI's messages API format.
• Explicitly states the generative model used (OpenRouter gpt-5-chat).

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Data may reflect bias inherent to the generative model and its training data.

Provenance

Source

Aratako on Hugging Face.

Collection Method

Synthetically generated using the OpenRouter gpt-5-chat model.

Time Range

Creation date not specified; last updated August 19, 2025.

Freshness

Last updated 2025-08-19 07:11:43; freshness should be verified.

Geography

null