Hugging Face's logo

Datasets:
yatharth97
/
isa_gemma

Modalities:
Text
Formats:
csv
Size:
1K - 10K
Libraries:
Datasets
pandas
Dataset card Viewer Files Community
1
isa_gemma / README.md
yatharth97's picture
yatharth97
Create README.md
1e00bca verified
preview code
|
raw
history blame contribute delete
2.79 kB

Dataset Card for Company-Specific Financial Analysis Dataset

Dataset Description

This dataset contains structured conversation data formatted to train and evaluate natural language processing models on tasks related to financial analysis of specific companies. Each entry in the dataset consists of a conversation ID, steps within the conversation, system prompts, user questions, and the corresponding model responses.

Fields Overview

  • conv_id: Unique identifier for each conversation.
  • step: Sequence number within the conversation indicating the order of interaction.
  • system_prompt: Instructions or contextual information provided to guide the conversation.
  • question: User-generated query related to specific financial aspects of a company.
  • response: Model-generated response corresponding to the user's query.

Intended Use

This dataset is intended to develop and evaluate conversational agents capable of conducting informed discussions about company-specific financial matters such as stock performance, latest news, investment risks, and ESG ratings.

Limitations

The dataset is narrowly focused on financial topics and may not generalize well to other domains. Additionally, it is tailored to the specifics of conversational models, which may not be directly applicable to other types of language processing tasks.

Dataset Size and Structure

  • Number of rows: 2.69k
  • Example Entry:
    • conv_id: 0
    • step: stock_info
    • system_prompt: As a highly intelligent assistant and successor of...
    • question: Please summarize the stock performance of Boeing during...
    • response: stock_info:Boeing:6m

Data Fields

  • conv_id (int64): A unique integer identifying each conversation thread.
  • step (string): Labels describing the context or topic of the conversation step.
  • system_prompt (string): Predefined prompts that set the stage for the AI's response.
  • question (string): Queries posed by users, focused on specific aspects of a company's financial health.
  • response (string): Contains tagged responses where each tag identifies the type of information followed by the specific content (e.g., stock_info:Boeing:6m).

Source

This dataset was compiled from simulated dialogues based on financial data sources, structured to mimic real-world conversations between financial analysts and an AI assistant.

Usage

The dataset is ideal for training conversational AI models to handle specific financial queries, providing responses that are informative and contextually relevant to the questions asked. It can be used to improve the natural language understanding capabilities of AI in the finance sector, especially for tasks involving detailed financial discussions.