# Natural Language Inference on MultiNLI mismatched (cross-domain) RepEval 2017 (test)

Source: [Natural Language Inference on MultiNLI mismatched (cross-domain) RepEval 2017 (test)](<https://www.sota2.com/research/sota/natural-language-inference-on-multinli-mismatched-cross-domain-repeval-2017-test>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-02-26T05:27:22.445Z

Evaluation of Natural Language Inference performance on the mismatched (cross-domain) test split of the MultiNLI dataset, typically measured by accuracy.

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Natural Language Inference |
| Dataset (including any stated split) | MultiNLI mismatched (cross-domain) RepEval 2017 (test) |
| Metric | Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 25 of 25 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [Accuracy](<https://www.sota2.com/research/sota/natural-language-inference-on-multinli-mismatched-cross-domain-repeval-2017-test.md?metric=Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| Our ESIM | 75.8 | Not recorded | 68fa638b1e209131ce9ff741 |
| Ensembled | 74.9 | Not recorded | 68fa638b1e209131ce9ff741 |
| Single | 73.6 | Not recorded | 68fa638b1e209131ce9ff741 |
| Single (Input Gate) | 73.6 | Not recorded | 68fa638b1e209131ce9ff741 |
| Single - Gated-Att | 73.6 | Not recorded | 68fa638b1e209131ce9ff741 |
| Shortcut-Stacked Encoder | 73.6 | Not recorded | 68fa63d11e209131cea00ec6 |
| YixinNie-UNC-NLP | 73.5 | Not recorded | 68fa638b1e209131ce9ff741 |
| Single - CharCNN | 73.5 | Not recorded | 68fa638b1e209131ce9ff741 |
| Single (Output Gate) | 73.4 | Not recorded | 68fa638b1e209131ce9ff741 |
| Single (Forget Gate) | 73.1 | Not recorded | 68fa638b1e209131ce9ff741 |
| Rivercorners | 72.8 | Not recorded | 68fa638b1e209131ce9ff741 |
| ESIM | 72.1 | Not recorded | 68fa638b1e209131ce9ff741 |
| Rivercorners | 72.1 | Not recorded | 68fa638b1e209131ce9ff741 |
| DiSAN | 71.402 | Not recorded | 68fa63cf1e209131cea00caa |
| biLSTM-Max Encoder | 71.2 | Not recorded | 68fa63d11e209131cea00ec6 |
| LCT-MALTA | 70.8 | Not recorded | 68fa638b1e209131ce9ff741 |
| Single - AbsDiff/Product | 69.2 | Not recorded | 68fa638b1e209131ce9ff741 |
| TALP-UPC | 68.2 | Not recorded | 68fa638b1e209131ce9ff741 |
| Bi-LSTM | 67.248 | Not recorded | 68fa63cf1e209131cea00caa |
| biLSTM Encoder | 67.1 | Not recorded | 68fa63d11e209131cea00ec6 |
| BiLSTM | 66.9 | Not recorded | 68fa638b1e209131ce9ff741 |
| Single - Word Embedding | 66 | Not recorded | 68fa638b1e209131ce9ff741 |
| CBOW | 64.759 | Not recorded | 68fa63cf1e209131cea00caa |
| CBOW | 64.6 | Not recorded | 68fa63d11e209131cea00ec6 |
| CBOW | 64.5 | Not recorded | 68fa638b1e209131ce9ff741 |

### Our ESIM

Accuracy: 75.8. Result ID: 698bfb81f729bc7f68e73459.

[Reporting paper](<https://www.sota2.com/research/paper/68fa638b1e209131ce9ff741.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67231e209131cea1499e.md>)

[Implementation](<https://github.com/lukecq1231/nli>)

Recorded evaluation configuration: {"implementation":"authors"}. This record does not assert independent reproduction.

### Ensembled

Accuracy: 74.9. Result ID: 698bfb81f729bc7f68e7345b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa638b1e209131ce9ff741.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa638b1e209131ce9ff741.md>)

[Implementation](<https://github.com/lukecq1231/enc_nli>)

Recorded evaluation configuration: {"ensemble":true,"RepEval 2017 participant":true}. This record does not assert independent reproduction.

### Single

Accuracy: 73.6. Result ID: 698bfb81f729bc7f68e7345a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa638b1e209131ce9ff741.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa638b1e209131ce9ff741.md>)

[Implementation](<https://github.com/lukecq1231/enc_nli>)

Recorded evaluation configuration: {"ensemble":false,"RepEval 2017 participant":true}. This record does not assert independent reproduction.

### Single (Input Gate)

Accuracy: 73.6. Result ID: 698bfb81f729bc7f68e7345c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa638b1e209131ce9ff741.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa638b1e209131ce9ff741.md>)

[Implementation](<https://github.com/lukecq1231/enc_nli>)

Recorded evaluation configuration: {"gate\_type":"input","RepEval 2017 participant":true}. This record does not assert independent reproduction.

### Single - Gated-Att

Accuracy: 73.6. Result ID: 698bfb81f729bc7f68e7345f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa638b1e209131ce9ff741.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa638b1e209131ce9ff741.md>)

[Implementation](<https://github.com/lukecq1231/enc_nli>)

Recorded evaluation configuration: {"gated\_attention":"removed"}. This record does not assert independent reproduction.

### Shortcut-Stacked Encoder

Accuracy: 73.6. Result ID: 698c798d37894ef7d9cea7c2.

[Reporting paper](<https://www.sota2.com/research/paper/68fa63d11e209131cea00ec6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa63d11e209131cea00ec6.md>)

[Implementation](<https://github.com/easonnie/multiNLI_encoder>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### YixinNie-UNC-NLP

Accuracy: 73.5. Result ID: 698bfb81f729bc7f68e73458.

[Reporting paper](<https://www.sota2.com/research/paper/68fa638b1e209131ce9ff741.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa63821e209131ce9ff0a5.md>)

Recorded evaluation configuration: {"RepEval 2017 participant":true}. This record does not assert independent reproduction.

### Single - CharCNN

Accuracy: 73.5. Result ID: 698bfb81f729bc7f68e73460.

[Reporting paper](<https://www.sota2.com/research/paper/68fa638b1e209131ce9ff741.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa638b1e209131ce9ff741.md>)

[Implementation](<https://github.com/lukecq1231/enc_nli>)

Recorded evaluation configuration: {"character\_composition":"removed"}. This record does not assert independent reproduction.

### Single (Output Gate)

Accuracy: 73.4. Result ID: 698bfb81f729bc7f68e7345e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa638b1e209131ce9ff741.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa638b1e209131ce9ff741.md>)

[Implementation](<https://github.com/lukecq1231/enc_nli>)

Recorded evaluation configuration: {"gate\_type":"output"}. This record does not assert independent reproduction.

### Single (Forget Gate)

Accuracy: 73.1. Result ID: 698bfb81f729bc7f68e7345d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa638b1e209131ce9ff741.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa638b1e209131ce9ff741.md>)

[Implementation](<https://github.com/lukecq1231/enc_nli>)

Recorded evaluation configuration: {"gate\_type":"forget"}. This record does not assert independent reproduction.

### Rivercorners

Accuracy: 72.8. Result ID: 698bfb81f729bc7f68e73457.

[Reporting paper](<https://www.sota2.com/research/paper/68fa638b1e209131ce9ff741.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa63821e209131ce9ff0a5.md>)

Recorded evaluation configuration: {"RepEval 2017 participant":true,"ensemble":true}. This record does not assert independent reproduction.

### ESIM

Accuracy: 72.1. Result ID: 698bfb81f729bc7f68e73453.

[Reporting paper](<https://www.sota2.com/research/paper/68fa638b1e209131ce9ff741.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67231e209131cea1499e.md>)

[Implementation](<https://github.com/lukecq1231/nli>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Rivercorners

Accuracy: 72.1. Result ID: 698bfb81f729bc7f68e73456.

[Reporting paper](<https://www.sota2.com/research/paper/68fa638b1e209131ce9ff741.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa63821e209131ce9ff0a5.md>)

Recorded evaluation configuration: {"RepEval 2017 participant":true,"ensemble":false}. This record does not assert independent reproduction.

### DiSAN

Accuracy: 71.402. Result ID: 698c793137894ef7d9cea5e4.

[Reporting paper](<https://www.sota2.com/research/paper/68fa63cf1e209131cea00caa.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa63cf1e209131cea00caa.md>)

[Implementation](<https://github.com/taoshen58/DiSAN>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### biLSTM-Max Encoder

Accuracy: 71.2. Result ID: 698c798d37894ef7d9cea7c1.

[Reporting paper](<https://www.sota2.com/research/paper/68fa63d11e209131cea00ec6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa64851e209131cea051f9.md>)

[Implementation](<https://www.github.com/facebookresearch/InferSent>)

Recorded evaluation configuration: {"reimplementation":true}. This record does not assert independent reproduction.

### LCT-MALTA

Accuracy: 70.8. Result ID: 698bfb81f729bc7f68e73455.

[Reporting paper](<https://www.sota2.com/research/paper/68fa638b1e209131ce9ff741.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa63821e209131ce9ff0a5.md>)

Recorded evaluation configuration: {"RepEval 2017 participant":true}. This record does not assert independent reproduction.

### Single - AbsDiff/Product

Accuracy: 69.2. Result ID: 698bfb81f729bc7f68e73462.

[Reporting paper](<https://www.sota2.com/research/paper/68fa638b1e209131ce9ff741.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa638b1e209131ce9ff741.md>)

[Implementation](<https://github.com/lukecq1231/enc_nli>)

Recorded evaluation configuration: {"elementwise\_vector\_comparison":"removed"}. This record does not assert independent reproduction.

### TALP-UPC

Accuracy: 68.2. Result ID: 698bfb81f729bc7f68e73454.

[Reporting paper](<https://www.sota2.com/research/paper/68fa638b1e209131ce9ff741.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa63821e209131ce9ff0a5.md>)

Recorded evaluation configuration: {"RepEval 2017 participant":true}. This record does not assert independent reproduction.

### Bi-LSTM

Accuracy: 67.248. Result ID: 698c793137894ef7d9cea5e3.

[Reporting paper](<https://www.sota2.com/research/paper/68fa63cf1e209131cea00caa.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa62151e209131ce9f7669.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### biLSTM Encoder

Accuracy: 67.1. Result ID: 698c798d37894ef7d9cea7c0.

[Reporting paper](<https://www.sota2.com/research/paper/68fa63d11e209131cea00ec6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa64101e209131cea02761.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### BiLSTM

Accuracy: 66.9. Result ID: 698bfb81f729bc7f68e73452.

[Reporting paper](<https://www.sota2.com/research/paper/68fa638b1e209131ce9ff741.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa64101e209131cea02761.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Single - Word Embedding

Accuracy: 66. Result ID: 698bfb81f729bc7f68e73461.

[Reporting paper](<https://www.sota2.com/research/paper/68fa638b1e209131ce9ff741.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa638b1e209131ce9ff741.md>)

[Implementation](<https://github.com/lukecq1231/enc_nli>)

Recorded evaluation configuration: {"word\_level\_embedding":"removed"}. This record does not assert independent reproduction.

### CBOW

Accuracy: 64.759. Result ID: 698c793137894ef7d9cea5e2.

[Reporting paper](<https://www.sota2.com/research/paper/68fa63cf1e209131cea00caa.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa610a1e209131ce9f246e.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### CBOW

Accuracy: 64.6. Result ID: 698c798d37894ef7d9cea7bf.

[Reporting paper](<https://www.sota2.com/research/paper/68fa63d11e209131cea00ec6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa64101e209131cea02761.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### CBOW

Accuracy: 64.5. Result ID: 698bfb81f729bc7f68e73451.

[Reporting paper](<https://www.sota2.com/research/paper/68fa638b1e209131ce9ff741.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa64101e209131cea02761.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.
