| --- |
| pretty_name: AI Prompt & Conversation Dataset |
| tags: |
| - ai |
| - prompt-dataset |
| - conversation-dataset |
| - llm |
| - conversational-ai |
| - human-ai-interaction |
| - chat-data |
| - instruction-following |
| - natural-language-processing |
| - multimodal-ai |
| - machine-learning |
| - json |
| - text |
| - image |
| - pdf |
| --- |
| |
| # AI Prompt & Conversation Dataset |
|
|
| ## Overview |
|
|
| The **AI Prompt & Conversation Dataset** is a collection of user prompts and AI-generated conversations designed to represent interactions between people and AI assistants across a variety of topics and use cases. |
|
|
| The dataset contains conversation sessions with user messages, assistant responses, session-level summaries, and associated attachment records. It captures different types of interactions, including informational queries, technical discussions, educational questions, research-related conversations, recommendations, troubleshooting, and everyday assistance. |
|
|
| The collection can support the development, evaluation, and research of Large Language Models (LLMs), conversational AI systems, prompt engineering techniques, and AI-powered applications. |
|
|
| ## Dataset Highlights |
|
|
| - **Dataset Type:** AI prompts and conversations |
| - **Primary Content:** User messages and AI assistant responses |
| - **Organization:** Session-based conversation records |
| - **Conversation Information:** Session identifiers, tags, summaries, and message counts |
| - **Attachment Information:** Attachment identifiers, content types, file sizes, timestamps, and available image descriptions |
| - **Media Types Observed:** Images, PDFs, and plain-text files |
| - **Potential Applications:** LLM development, conversational AI, NLP, and AI system evaluation |
|
|
| ## Dataset Features |
|
|
| - **Multi-session conversations:** Conversations are organized into individual sessions. |
| - **User and assistant messages:** Contains exchanges between users and AI assistants. |
| - **Session summaries:** Includes summaries describing user intent, discussion topics, decisions, and relevant details. |
| - **Session metadata:** Includes identifiers, tags, version information, and message counts. |
| - **Attachment records:** Provides information about files associated with conversations. |
| - **Multimedia references:** Includes attachment content types and image descriptions where available. |
| - **Conversation diversity:** Covers technical, educational, informational, research, and everyday assistance scenarios. |
|
|
| ## Conversation Coverage |
|
|
| The sample workbook demonstrates conversations involving several categories: |
|
|
| - Artificial Intelligence and Large Language Models |
| - Software Development and Programming |
| - Technical Troubleshooting |
| - Education and Learning |
| - Research and Academic Discussions |
| - Business and Financial Topics |
| - Product and Service Recommendations |
| - Travel and General Information |
| - Sports and Entertainment |
| - Everyday Questions and Assistance |
|
|
|
|
| ## Dataset Structure |
|
|
| The supplied Excel workbook contains two worksheets: `sessions` and `attachments`. |
|
|
| ### 1. Sessions |
|
|
| The `sessions` worksheet contains conversation-level information and message content. |
|
|
| | Field | Description | |
| |---|---| |
| | `session_id` | Identifier for a conversation session | |
| | `session_tag` | Session title or topic label | |
| | `session_summary` | Summary of the conversation | |
| | `version_id` | Identifier associated with a conversation version | |
| | `version_number` | Version number, where available | |
| | `extracted_from` | Source or extraction information | |
| | `message_count` | Number of messages in the session | |
| | `messages` | Conversation messages and content | |
| | `attachment_count` | Number of associated attachments | |
|
|
| ### 2. Attachments |
|
|
| The `attachments` worksheet contains metadata about files associated with conversation sessions. |
|
|
| | Field | Description | |
| |---|---| |
| | `session_id` | Identifier of the associated session | |
| | `version_id` | Associated conversation version | |
| | `attachment_id` | Unique attachment identifier | |
| | `actual_content_type` | File content type or MIME type | |
| | `size_bytes` | Attachment size in bytes | |
| | `uploaded_at` | Attachment upload timestamp | |
| | `vision_caption` | Image description, where available | |
|
|
| ## Supported Content and File Types |
|
|
| The supplied sample workbook contains attachment records with the following content types: |
|
|
| - **JPEG:** `image/jpeg` |
| - **PNG:** `image/png` |
| - **PDF:** `application/pdf` |
| - **Plain Text:** `text/plain` |
|
|
| The dataset also contains textual conversation content, including user prompts, assistant responses, and session summaries. |
|
|
| ## AI Tasks Supported |
|
|
| The dataset may be useful for research and development involving: |
|
|
| - Large Language Model (LLM) Development |
| - Conversational AI Training |
| - Prompt Engineering and Analysis |
| - Human-AI Interaction Research |
| - Natural Language Processing (NLP) |
| - Instruction Following |
| - Conversation Summarization |
| - Intent and Topic Classification |
| - Context-Aware Response Generation |
| - Multimodal AI Research |
| - AI Assistant Evaluation |
|
|
| ## Potential Applications |
|
|
| - Training and evaluating conversational AI systems. |
| - Developing AI-powered chatbots and virtual assistants. |
| - Researching human-AI interaction patterns. |
| - Studying prompt structures and conversation flows. |
| - Developing conversation summarization systems. |
| - Evaluating AI response generation and instruction-following capabilities. |
| - Building systems that work with conversational and attachment metadata. |
| - Researching multimodal conversational applications. |
|
|
| ## Industries and Research Areas |
|
|
| Potential areas of application include: |
|
|
| - Artificial Intelligence and Machine Learning |
| - Generative AI |
| - Enterprise AI Solutions |
| - Software Development |
| - Education Technology |
| - Natural Language Processing |
| - Conversational AI |
| - Customer Support Automation |
| - AI Research and Development |
| - Digital Assistant Platforms |
|
|
| ## Intended Audience |
|
|
| This dataset may be useful for: |
|
|
| - AI and machine learning researchers |
| - LLM developers |
| - Generative AI companies |
| - Conversational AI developers |
| - NLP researchers |
| - AI application developers |
| - Enterprise AI teams |
| - Academic institutions |
| - Data scientists |
| - AI evaluation and benchmarking teams |
|
|
| ## Dataset Statistics |
|
|
| The following figures are taken from the supplied **AI Prompt & Conversation Datasets Catalog** screenshot. They represent catalogue-level statistics and should not be interpreted as counts from the sample Excel workbook alone. |
|
|
| | Metric | Catalogue figure | |
| |---|---:| |
| | Sessions | 259,198 | |
| | Messages | 13.4M | |
| | Estimated Tokens | 2.75B | |
| | Attachments | 607,049 | |
| | Media Payload | 50.59 GB | |
| | Reuse Events | 200,629 | |
|
|
| ### Sample Workbook Statistics |
|
|
| The separately supplied `DEMOdata.xlsx` workbook contains: |
|
|
| | Metric | Sample figure | |
| |---|---:| |
| | Conversation sessions | 52 | |
| | Messages | 752 | |
| | Attachment records | 27 | |
|
|
| The sample attachment records include 18 JPEG files, 3 PNG files, 4 PDF files, and 2 plain-text files. |
|
|
| ## Advantages |
|
|
| - Provides examples of multi-turn interactions between users and AI assistants. |
| - Includes both conversation content and session-level metadata. |
| - Offers summaries that help describe conversation context and user intent. |
| - Includes attachment metadata for conversations involving additional files. |
| - Supports research into different conversational topics and use cases. |
| - Provides a basis for studying AI interaction patterns and conversational workflows. |
|
|
|
|
| ## Summary |
|
|
| The **AI Prompt & Conversation Dataset** provides a collection of AI-user interactions organized into conversation sessions, with message content, session summaries, metadata, and associated attachment information. With catalogue-level figures reporting 259,198 sessions, 13.4 million messages, and 2.75 billion estimated tokens, the collection may be valuable for LLM research, conversational AI development, prompt analysis, and human-AI interaction studies. |
|
|
| **Data Creation** |
|
|
| Procured through formal agreements and generated in the ordinary course of business. |
|
|
| **Considerations** |
|
|
| This dataset is provided for research and educational purposes only. It contains only sample data. For access to the full dataset and enterprise licensing options, please visit our website [InfoBay.AI](https://infobay.ai/) or contact us directly. |
|
|
| -Ph: +91 9236398619 |
| -Email: datareq@infobay.ai |
| |
