ToolBunny Logo
ToolBunny
Back to Homepage

CSV Few-Shot Dataset Factory

Format explicit user and system message pairs smoothly to construct LLM dataset files.

CSV / TSV Dataset File Source:
3 rows parsed
Interactive Column Mapping
Dataset Code Output:
{"messages":[{"role":"system","content":"You are a helpful customer support bot."},{"role":"user","content":"How do I reset my password?"},{"role":"assistant","content":"Click on 'Forgot Password' on the login screen and follow the email link."}]}
{"messages":[{"role":"system","content":"You are a helpful customer support bot."},{"role":"user","content":"Where is my order tracking number?"},{"role":"assistant","content":"Log in to your account, navigate to 'My Orders', and select your shipment."}]}
{"messages":[{"role":"system","content":"You are a helpful customer support bot."},{"role":"user","content":"Can I get a refund?"},{"role":"assistant","content":"Refunds are processed within 14 days of purchase. Please contact support@example.com."}]}
Dataset Pairs Formatted
3 Rows
Est. Total Token Volume
~199 Tokens

About This Tool

The CSV Few-Shot Dataset Factory is an essential data preparation utility engineered for prompt engineers, AI developers, and machine learning researchers preparing training datasets and few-shot prompt blocks.

Converting spreadsheets, CSV exports, or tabular user/assistant conversation logs into valid OpenAI fine-tuning .jsonl files or formatted few-shot prompt blocks usually requires writing custom Python or Pandas scripts. This tool parses raw CSV files, maps columns interactively to user/assistant/system roles, applies global system prompts, and exports ready-to-use dataset files.

How to Use

  1. Upload or Paste CSV: Drag and drop a CSV/TSV file or paste tabular text into the CSV editor.
  2. Map Columns Interactively: Select which CSV column represents the User Prompt, Assistant Response, and System Instruction.
  3. Prepend Global System Instruction (Optional): Enter a universal system prompt to apply across all conversation turns.
  4. Choose Target Dataset Format: Switch between OpenAI JSONL, HuggingFace JSONL, Alpaca JSON, or Few-Shot Prompt Text format.
  5. Copy or Download File: Copy the output code directly to your clipboard or download a .jsonl, .json, or .txt file.

Common Use Cases & FAQ