ConditionalNLI/PROVISER

Proviso-NLI Dataset

A Natural Language Inference (NLI) dataset with examples constructed using different trigger types.

Dataset Summary

This dataset contains Natural Language Inference examples across three subsets (2, 3, and 4) with three different trigger types:

• Type 4: Possessive triggers
• Type 5a: Again triggers
• Type 5p: Possessive triggers

The dataset is designed for evaluating NLI models.

Dataset Structure

Data Fields

• premise (string): The premise sentence
• hypothesis (string): The hypothesis sentence
• gold_label (string): Gold standard label (E, C, or N for Entailment, Contradiction, Neutral)
• trigger (string): The type of trigger used (e.g., "possessive", "again")
• SetID (int): Set identifier
• UID (string): Unique identifier
• trigger_words (string): The specific trigger words in the example
• premise_type (string): The type of premise (e.g., "Attitude Premise", "Conjunction Premise")

Splits

The dataset is organized into 9 subsets:

Subset	Type	Description
subset2_type4	Possessive	Subset 2 - Possessive triggers
subset2_type5a	Again	Subset 2 - Again triggers
subset2_type5p	Possessive	Subset 2 - Possessive triggers
subset3_type4	Possessive	Subset 3 - Possessive triggers
subset3_type5a	Again	Subset 3 - Again triggers
subset3_type5p	Possessive	Subset 3 - Possessive triggers
subset4_type4	Possessive	Subset 4 - Possessive triggers
subset4_type5a	Again	Subset 4 - Again triggers
subset4_type5p	Possessive	Subset 4 - Possessive triggers

Loading the Dataset

from datasets import load_dataset

# Load a specific subset
dataset = load_dataset("ConditionalNLI/PROVISER", name="subset2_type4")

# Load all subsets
datasets = load_dataset("ConditionalNLI/PROVISER")