bonmuq/kms_broadcast_eval
Dataset: KMS Broadcast Eval Description This dataset contains financial broadcast evaluation data for compliance checks. The dataset includes labeled data to identify violations based on specific criteria. Files kms_broadcast_eval.csv: The main dataset containing context, violation status, reasons, and categories. Label Definitions 0: Not violated 1: Violated Columns id: Unique identifier for each row context: The… See the full description on the dataset page: https://huggingface.co/datasets/bonmuq/kms_broadcast_eval.
Dataset: KMS Broadcast Eval
Description
This dataset contains financial broadcast evaluation data for compliance checks. The dataset includes labeled data to identify violations based on specific criteria.
Files
- kms_broadcast_eval.csv: The main dataset containing context, violation status, reasons, and categories.
Label Definitions
- 0: Not violated
- 1: Violated
Columns
- id: Unique identifier for each row
- context: The actual advertisement text that is being evaluated
- violation_status: Indicates whether there is a violation in the given context (0 for not violated, 1 for violated)
- violation_reason: The reasons for the violation, listing the specific phrases that are problematic
- category_1: Violation status for "단정적인 표현" or "절대적 수익" (0 for not violated, 1 for violated)
- category_2: Violation status for "손실 보전" or "수익 보장" (0 for not violated, 1 for violated)
- category_3: Violation status for "근거없거나 과도한 최상급 혹은 긍정적 표현" (0 for not violated, 1 for violated)
- category_4: Violation status for "수익률 관련 금지표현 및 권장표현" (0 for not violated, 1 for violated)
- category_5: Violation status for "금융 상품 간의 비교 표현" (0 for not violated, 1 for violated)