19 datasets found

A new dataset for handling entailments with lexical and logical phenomena, called HELP, is introduced to investigate the shortcomings of neural models in monotonicity reasoning.

The BoolQ dataset is a context-based yes-no question answering dataset.

The dataset used in the paper is SNLI, SciTail, and BreakingNLI.

SNLI, MNLI, Stress Test, Matched Mismatched, Competence, Distraction, and Noise datasets

KLUE-NLI dataset contains sentence pairs from Korean language

The dataset used for Natural Language Inference task, which involves a premise P = {p0, p1,..., pm−1} of m words and a hypothesis H = {h0, h1,..., hn−1} of n words, and aims to...

The Stanford Corpus of Implicatives is a dataset of natural language inference tasks.

SICK-NL is a dataset for Dutch Natural Language Inference, obtained by translating the SICK dataset from English into Dutch.

The dataset used for investigating the extent to which sentence representations arising from neural machine translation (NMT) systems encode distinct semantic phenomena.

The paper proposes a novel approach to the sequential example selection paradigm for in-context learning.

The e-SNLI dataset is a large-scale dataset for natural language inference, with human-written explanations for each sample.

The dataset used in the paper is the SNLI dataset.

The MNLI and FEVER datasets are used to evaluate the proposed MoCaD framework.

The dataset is used for instruction-tuning of LLMs in multiple languages using reinforcement learning from human feedback.

The Breaking NLI dataset is a test set constructed by taking premises from the SNLI training set and constructing several hypotheses from them by changing at most one word...

The XNLI dataset comprises pairs of sentences with a label categorizing the semantic relationship between the two sentences into one of three classifications: entailment,...

CIRR is a general image dataset that comprises 36,554 triplets derived from 21,552 images from the popular natural language inference dataset NLVR2.

The dataset used in the paper is the Stanford Natural Language Inference (SNLI) dataset, which consists of 549,367 premise-hypothesis pairs for train/dev/test sets and target...

SICK is a dataset for recognizing textual entailment (RTE), containing 4.5K/0.5K/5.0K train/dev/test examples. Each example consists of a hypothesis and a premise, and the goal...