SOTAVerified

Natural Language Inference

Natural language inference (NLI) is the task of determining whether a "hypothesis" is true (entailment), false (contradiction), or undetermined (neutral) given a "premise".

Example:

| Premise | Label | Hypothesis | | --- | ---| --- | | A man inspects the uniform of a figure in some East Asian country. | contradiction | The man is sleeping. | | An older and younger man smiling. | neutral | Two men are smiling and laughing at the cats playing on the floor. | | A soccer game with multiple males playing. | entailment | Some men are playing a sport. |

Approaches used for NLI include earlier symbolic and statistical approaches to more recent deep learning approaches. Benchmark datasets used for NLI include SNLI, MultiNLI, SciTail, among others. You can get hands-on practice on the SNLI task by following this d2l.ai chapter.

Further readings:

Papers

Showing 110 of 1961 papers

TitleStatusHype
LRCTI: A Large Language Model-Based Framework for Multi-Step Evidence Retrieval and Reasoning in Cyber Threat Intelligence Credibility Verification0
DS@GT at CheckThat! 2025: Evaluating Context and Tokenization Strategies for Numerical Fact VerificationCode0
ARAG: Agentic Retrieval Augmented Generation for Personalized Recommendation0
When Does Meaning Backfire? Investigating the Role of AMRs in NLI0
Thunder-NUBench: A Benchmark for LLMs' Sentence-Level Negation Understanding0
Explainable Compliance Detection with Multi-Hop Natural Language Inference on Assurance Case Structure0
Theorem-of-Thought: A Multi-Agent Framework for Abductive, Deductive, and Inductive Reasoning in Language ModelsCode0
A MISMATCHED Benchmark for Scientific Natural Language InferenceCode0
CLATTER: Comprehensive Entailment Reasoning for Hallucination Detection0
Drop Dropout on Single-Epoch Language Model PretrainingCode0
Show:102550
← PrevPage 1 of 197Next →

Benchmark Results

#ModelMetricClaimedVerifiedStatus
1UnitedSynT5 (3B)% Test Accuracy94.7Unverified
2UnitedSynT5 (335M)% Test Accuracy93.5Unverified
3EFL (Entailment as Few-shot Learner) + RoBERTa-large% Test Accuracy93.1Unverified
4Neural Tree Indexers for Text Understanding% Test Accuracy93.1Unverified
5RoBERTa-large+Self-Explaining% Test Accuracy92.3Unverified
6RoBERTa-large + self-explaining layer% Test Accuracy92.3Unverified
7CA-MTL% Test Accuracy92.1Unverified
8SemBERT% Test Accuracy91.9Unverified
9MT-DNN-SMARTLARGEv0% Test Accuracy91.7Unverified
10MT-DNN-SMART_100%ofTrainingDataDev Accuracy91.6Unverified
#ModelMetricClaimedVerifiedStatus
1Vega v2 6B (KD-based prompt transfer)Accuracy96Unverified
2PaLM 540B (fine-tuned)Accuracy95.7Unverified
3Turing NLR v5 XXL 5.4B (fine-tuned)Accuracy94.1Unverified
4ST-MoE-32B 269B (fine-tuned)Accuracy93.5Unverified
5DeBERTa-1.5BAccuracy93.2Unverified
6MUPPET Roberta LargeAccuracy92.8Unverified
7DeBERTaV3largeAccuracy92.7Unverified
8T5-XXL 11BAccuracy92.5Unverified
9T5-XXL 11B (fine-tuned)Accuracy92.5Unverified
10ST-MoE-L 4.1B (fine-tuned)Accuracy92.1Unverified
#ModelMetricClaimedVerifiedStatus
1UnitedSynT5 (3B)Matched92.6Unverified
2Turing NLR v5 XXL 5.4B (fine-tuned)Matched92.6Unverified
3T5-XXL 11B (fine-tuned)Matched92Unverified
4T5Matched92Unverified
5T5-11BMismatched91.7Unverified
6T5-3BMatched91.4Unverified
7ALBERTMatched91.3Unverified
8DeBERTa (large)Matched91.1Unverified
9Adv-RoBERTa ensembleMatched91.1Unverified
10SMARTRoBERTaDev Matched91.1Unverified