Browse State-of-the-Art › Natural Language Inference

Natural Language Inference

821 papers with code · 33 benchmarks · 81 datasets archive 2025-07-28

Natural Language ProcessingReasoning

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:

Description from the archive archive 2025-07-28; Papers-with-Code links inside it are rewritten to this site.

Benchmarks archive 2025-07-28

37 leaderboard tables shown for this task, 33 with rows (a “benchmark” on this site is a table with at least one row, as on /sota), ordered by row count. “Best model” is the first row in the archive's own order at snapshot; nothing is re-ranked here and metric direction is not recorded in the archive. PwC's Trend sparklines are not in the archive, so that column is omitted. 10 shown of 37 until expanded.

Syntology 41 newer papers were judged to report on this task's benchmarks so far (Claude Sonnet 4.5, reading their tables: a model's judgement, not a result; 1 placed): SNLI (4 judged, 0 placed), RTE (28 judged, 1 placed), MultiNLI (18 judged, 0 placed), QNLI (21 judged, 1 placed), WNLI (1 judged, 0 placed), SciTail (1 judged, 0 placed), MultiNLI Dev (3 judged, 0 placed), V-SNLI (1 judged, 0 placed), MRPC (2 judged, 0 placed), TabFact (1 judged, 0 placed). Drawn from 8,886 of 18,637 newer papers checked so far (733 more had nothing to place; 41 could not be read).

DatasetBest model (first row in archive order)PaperCodeSyntologyCompare
SNLI (98 rows) UnitedSynT5 (3B) First Train to Generate, then Generate to Train: UnitedSynT5 for... — — Compare
RTE (90 rows) Vega v2 6B (KD-based prompt transfer) Toward Efficient Language Model Pretraining and Downstream... — — Compare
MultiNLI (67 rows) Turing NLR v5 XXL 5.4B (fine-tuned) — — — Compare
QNLI (43 rows) ALBERT ALBERT: A Lite BERT for Self-supervised Learning of Language... code Syntology official (archive's flag): 6 ran · 81 ran (of which 17 constructed an object rather than computing a result; 56 with no instrument failure: 4 honoured, 0 violated, 52 with no contract checked; 25 where Syntology's instrument failed) · 45 unverified Compare
ANLI test (25 rows) T5-3B (explanation prompting) Prompting for explanations improves Adversarial NLI. Is this true?... — — Compare
WNLI (23 rows) Turing NLR v5 XXL 5.4B (fine-tuned) — — — Compare
LiDiRus (22 rows) Human Benchmark RussianSuperGLUE: A Russian Language Understanding Evaluation Benchmark code Syntology official (archive's flag): 1 ran · 1 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 1 honoured, 0 violated, 0 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified Compare
RCB (22 rows) Human Benchmark RussianSuperGLUE: A Russian Language Understanding Evaluation Benchmark code Syntology official (archive's flag): 1 ran · 1 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 1 honoured, 0 violated, 0 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified Compare
TERRa (22 rows) Human Benchmark RussianSuperGLUE: A Russian Language Understanding Evaluation Benchmark code Syntology official (archive's flag): 1 ran · 1 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 1 honoured, 0 violated, 0 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified Compare
CommitmentBank (20 rows) PaLM 540B (finetuned) PaLM: Scaling Language Modeling with Pathways code Syntology 32 ran (of which 16 constructed an object rather than computing a result; 24 with no instrument failure: 2 honoured, 1 violated, 21 with no contract checked; 8 where Syntology's instrument failed) · 5 unverified Compare
SciTail (13 rows) CA-MTL Conditionally Adaptive Multi-Task Learning: Improving Transfer... code — Compare
FarsTail (10 rows) mBERT FarsTail: A Persian Natural Language Inference Dataset code — Compare
MultiNLI Dev (10 rows) TinyBERT-6 67M TinyBERT: Distilling BERT for Natural Language Understanding code Syntology official: harvested, nothing ran · 0 ran · 4 unverified Compare
MedNLI (7 rows) ClinicalMosaic Patient Trajectory Prediction: Integrating Clinical Notes with Transformers code — Compare
XNLI French (6 rows) CamemBERT (large) CamemBERT: a Tasty French Language Model code — Compare
e-SNLI (3 rows) ExplainThenPredictAttention (e-InferSent Bi-LSTM + Attention) e-SNLI: Natural Language Inference with Natural Language Explanations code — Compare
V-SNLI (3 rows) MMBT Supervised Multimodal Bitransformers for Classifying Images and Text code — Compare
XNLI Chinese Dev (3 rows) ERNIE 2.0 Large ERNIE 2.0: A Continual Pre-training Framework for Language Understanding code Syntology official: harvested, nothing ran · 0 ran · 1 unverified Compare
XNLI Chinese (3 rows) ERNIE 2.0 Large ERNIE 2.0: A Continual Pre-training Framework for Language Understanding code Syntology official: harvested, nothing ran · 0 ran · 1 unverified Compare
JamPatoisNLI (2 rows) roberta-unfrozen JamPatoisNLI: A Jamaican Patois Natural Language Inference Dataset — — Compare
AX (1 row) T5 SMART: Robust and Efficient Fine-Tuning for Pre-trained Natural... code Syntology official: no sample here; runs from other or unrecorded repositories · 8 ran (of which 0 constructed an object rather than computing a result; 5 with no instrument failure: 2 honoured, 1 violated, 2 with no contract checked; 3 where Syntology's instrument failed) · 0 unverified Compare
BioNLI (1 row) BioLinkBert BioNLI: Generating a Biomedical NLI Dataset Using Lexico-semantic... code — Compare
HANS (1 row) Roberta-large Generating Data to Mitigate Spurious Correlations in Natural... code — Compare
KUAKE-QQR (1 row) BERT-base CBLUE: A Chinese Biomedical Language Understanding Evaluation Benchmark code Syntology official (archive's flag): 11 ran · 11 ran (of which 0 constructed an object rather than computing a result; 8 with no instrument failure: 1 honoured, 0 violated, 7 with no contract checked; 3 where Syntology's instrument failed) · 5 unverified Compare
KUAKE-QTR (1 row) MacBERT-large CBLUE: A Chinese Biomedical Language Understanding Evaluation Benchmark code Syntology official (archive's flag): 11 ran · 11 ran (of which 0 constructed an object rather than computing a result; 8 with no instrument failure: 1 honoured, 0 violated, 7 with no contract checked; 3 where Syntology's instrument failed) · 5 unverified Compare
MED (1 row) NeuralLog NeuralLog: Natural Language Inference with Joint Neural and... code Syntology official (archive's flag): 3 ran · 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 3 unverified Compare
MNLI + SNLI + ANLI + FEVER (1 row) SMARTRoBERTa-LARGE SMART: Robust and Efficient Fine-Tuning for Pre-trained Natural... code Syntology official: no sample here; runs from other or unrecorded repositories · 8 ran (of which 0 constructed an object rather than computing a result; 5 with no instrument failure: 2 honoured, 1 violated, 2 with no contract checked; 3 where Syntology's instrument failed) · 0 unverified Compare
MRPC (1 row) DeBERTaV3large DeBERTaV3: Improving DeBERTa using ELECTRA-Style Pre-Training with... code Syntology official (archive's flag): 6 ran · 6 ran (of which 0 constructed an object rather than computing a result; 6 with no instrument failure: 0 honoured, 0 violated, 6 with no contract checked; 0 where Syntology's instrument failed) · 1 unverified Compare
Probability words NLI (1 row) roberta-base-mnli Probing neural language models for understanding of words of... — — Compare
Quora Question Pairs (1 row) aESIM Attention Boosted Sequential Inference Model — — Compare
SICK (1 row) NeuralLog NeuralLog: Natural Language Inference with Joint Neural and... code Syntology official (archive's flag): 3 ran · 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 3 unverified Compare
TabFact (1 row) ChatGPT 3.5 SpatialFormat LAPDoc: Layout-Aware Prompting for Documents — — Compare
XWINO (1 row) mGPT mGPT: Few-Shot Learners Go Multilingual code — Compare
ANLI (0 rows) no rows in the archive — —
ANLI-r3 (0 rows) no rows in the archive — —
GLUE (0 rows) no rows in the archive — —
multi_nli (0 rows) no rows in the archive — —

Syntology column: samples harvested from the paper's repositories and executed on synthesized fixtures; “ran” is not a correctness claim and does not order the table. Each line reads “N ran (of which C constructed an object rather than computing a result; K with no instrument failure: H honoured, V violated, P with no contract checked; I where Syntology's instrument failed) · U unverified”; the instrument figure counts failures of Syntology's instrument, not of the code. When the archive marks a repository official for the row's paper, the line starts with that repository's state (“official” is the archive's flag, not a verdict on who wrote the code; “community repositories only” when every sample that ran came from a community repository, “official: no sample here; runs from other or unrecorded repositories” when some came from a repository the paper names or has in its text, or from none recorded); hover it for the repositories the samples that ran came from. A dash means no Syntology record for that paper, not a recorded non-run. Each “ran” line links to that paper's sample list. Syntology's record for this page last changed ; when this build read Syntology's graph is in the build record. For agents: get_harvested_code_for_paper(arxiv_id) lists each paper's samples; how to connect.

Libraries

Not in the archive: the export carries no per-task library table, so there is nothing to show at snapshot 2025-07-28.

Datasets archive 2025-07-28

81 datasets whose archive record lists this task, ordered by the archive's paper count. 30 shown of 81 until expanded.

Subtasks archive 2025-07-28

3 subtasks in the archive's task tree.

Papers archive 2025-07-28

archive papers tagged: 1,961 · with a code link: 821 · where Syntology ran a sample: 209 (167 with a run with no instrument failure, 42 where every run was a failure of Syntology's instrument) Syntology

Show: all tagged papersonly where code ran (209 of 1,961 tagged: 167 with a run with no instrument failure, 42 where every run was a failure of Syntology's instrument)

Page 1 of 20: papers 1 to 100 of 1,961, in archive order: by repositories listed in the archive (most first), then newest first, not by stars (the archive holds no stars, so PwC's “Social” and “Latest” sorts cannot be reproduced). Papers that list no repository come after every paper that lists one.

Papers without a page here are shown as plain text. A Syntology line reads “N ran (of which C constructed an object rather than computing a result; K with no instrument failure: H honoured, V violated, P with no contract checked; I where Syntology's instrument failed) · U unverified”; the figure “where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. When the archive marks a repository official for the paper, the line starts with that repository's state (the archive's flag, not a verdict on who wrote the code); hover it for the repositories the samples that ran came from. The first sentence of the abstract is shown for the first 30 papers; every abstract is on its paper's page.

Syntology lines on 70 of the papers shown; no Syntology record for the others (a paper without an arXiv id cannot be joined to the graph, and absence from the graph layer is not a recorded non-run). “Ran” means the sample executed on a synthesized fixture, not that the paper's result was reproduced; each line links to that paper's sample list. Syntology's record for this page last changed ; when this build read Syntology's graph is in the build record. For agents: get_harvested_code_for_paper(arxiv_id) lists each paper's samples; how to connect.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections