Create README.md

README.md

@@ -0,0 +1,58 @@
+
+---
+annotations_creators: [machine-generated, human-validated]
+language_creators: [crowdsourced]
+language: [en]
+license: cc-by-4.0
+multilinguality: [monolingual]
+pretty_name: HR Intent Classification Dataset
+size_categories: [n<1K]
+source_datasets: []
+task_categories: [text-classification]
+task_ids: [intent-classification]
+paperswithcode_id: null
+---
+
+# Dataset Card for HR Intent Dataset
+
+## Dataset Summary
+
+A dataset for intent classification in enterprise HR workflows. Each row contains a user query, context (as a structured field with domain, topic, subject), and a label for the HR intent.
+
+## Supported Tasks and Leaderboards
+
+- Intent Classification (text classification)
+- Suitable for benchmarking BERT, RoBERTa, etc., on real-world HR requests.
+
+## Languages
+
+English
+
+## Dataset Structure
+
+### Data Fields
+
+- `query`: The user's free-form text (string)
+- `context`: HR domain info (stringified dict/object)
+- `intent`: The label (string, e.g., "generate-offer" or "review-contract")
+
+### Example
+
+| query                              | context                                          | intent                  |
+|-------------------------------------|--------------------------------------------------|-------------------------|
+| Make an offer letter for Priya S... | {domain: HR, topic: onboarding, subject: offer}  | generate-offer          |
+
+
+## Uses
+
+- Intended for training and evaluating intent classifiers for enterprise chatbots or workflow automation.
+
+## Ethical Considerations, Limitations
+
+- Only English, only HR domain.
+- Class balance: check or describe distribution.
+- Not guaranteed to be bias-free; review intent coverage.
+
+
+
+---