# Figurative Language Detection on IRFL

Source: [Figurative Language Detection on IRFL](<https://www.sota2.com/research/sota/figurative-language-detection-on-irfl>)

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

Data updated: 2026-06-04T13:30:12.645Z

Evaluation of models on the IRFL dataset for detecting various forms of figurative language, including idioms, similes, and metaphors.

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 | Figurative Language Detection |
| Dataset (including any stated split) | IRFL |
| Metric | Idioms Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 15 of 15 |
| 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

- [Idioms Accuracy](<https://www.sota2.com/research/sota/figurative-language-detection-on-irfl.md?metric=Idioms%20Accuracy>): Higher is better; unit: PERCENT.

- [Simile Accuracy](<https://www.sota2.com/research/sota/figurative-language-detection-on-irfl.md?metric=Simile%20Accuracy>): Higher is better; unit: PERCENT.

- [Metaphor (OoD) Accuracy](<https://www.sota2.com/research/sota/figurative-language-detection-on-irfl.md?metric=Metaphor%20(OoD)%20Accuracy>): Higher is better; unit: PERCENT.

- [Overall Accuracy](<https://www.sota2.com/research/sota/figurative-language-detection-on-irfl.md?metric=Overall%20Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Idioms Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| GRU | 46.4 | Not recorded | 6a2106eec10484bac23f0669 |
| gMLP | 43.8 | Not recorded | 6a2106eec10484bac23f0669 |
| RePercENT | 42.5 | Not recorded | 6a2106eec10484bac23f0669 |
| CLIP | 38.1 | Not recorded | 6a2106eec10484bac23f0669 |
| GRU | 36.1 | Not recorded | 6a2106eec10484bac23f0669 |
| CLIP | 36 | Not recorded | 6a2106eec10484bac23f0669 |
| gMLP | 33.7 | Not recorded | 6a2106eec10484bac23f0669 |
| RePercENT | 33.6 | Not recorded | 6a2106eec10484bac23f0669 |
| RePercENT | 32.6 | Not recorded | 6a2106eec10484bac23f0669 |
| RePercENT | 30.2 | Not recorded | 6a2106eec10484bac23f0669 |
| RePercENT | 28.2 | Not recorded | 6a2106eec10484bac23f0669 |
| CLIP | 26 | Not recorded | 6a2106eec10484bac23f0669 |
| CLIP-ViT-B/32 | 22.5 | Not recorded | 6a2106eec10484bac23f0669 |
| CLIP-ViT-B/32 | 20.4 | Not recorded | 6a2106eec10484bac23f0669 |
| CLIP-ViT-B/32 | 16 | Not recorded | 6a2106eec10484bac23f0669 |

### GRU

Idioms Accuracy: 46.4. Result ID: 6a217dbffbc373592c1efe94.

[Reporting paper](<https://www.sota2.com/research/paper/6a2106eec10484bac23f0669.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa619e1e209131ce9f5331.md>)

[Implementation](<https://github.com/jych/librnn.git>)

Recorded evaluation configuration: {"Input Modalities":"Image vs Caption ⊕ Definition"}. This record does not assert independent reproduction.

### gMLP

Idioms Accuracy: 43.8. Result ID: 6a217dbffbc373592c1efe93.

[Reporting paper](<https://www.sota2.com/research/paper/6a2106eec10484bac23f0669.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a101e209131cea25ed6.md>)

Recorded evaluation configuration: {"Input Modalities":"Image vs Caption ⊕ Definition"}. This record does not assert independent reproduction.

### RePercENT

Idioms Accuracy: 42.5. Result ID: 6a217dbffbc373592c1efe95.

[Reporting paper](<https://www.sota2.com/research/paper/6a2106eec10484bac23f0669.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a2106eec10484bac23f0669.md>)

Recorded evaluation configuration: {"Input Modalities":"Image vs Caption ⊕ Definition"}. This record does not assert independent reproduction.

### CLIP

Idioms Accuracy: 38.1. Result ID: 6a217dbffbc373592c1efe92.

[Reporting paper](<https://www.sota2.com/research/paper/6a2106eec10484bac23f0669.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69581e209131cea21ad0.md>)

[Implementation](<https://github.com/OpenAI/CLIP>)

Recorded evaluation configuration: {"Backbone":"ViT-B/32","Protocol":"finetune-all","Input Modalities":"Image vs Caption ⊕ Definition"}. This record does not assert independent reproduction.

### GRU

Idioms Accuracy: 36.1. Result ID: 6a217dbffbc373592c1efea2.

[Reporting paper](<https://www.sota2.com/research/paper/6a2106eec10484bac23f0669.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa619e1e209131ce9f5331.md>)

[Implementation](<https://github.com/jych/librnn.git>)

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

### CLIP

Idioms Accuracy: 36. Result ID: 6a217dbffbc373592c1efe91.

[Reporting paper](<https://www.sota2.com/research/paper/6a2106eec10484bac23f0669.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69581e209131cea21ad0.md>)

[Implementation](<https://github.com/OpenAI/CLIP>)

Recorded evaluation configuration: {"Backbone":"ViT-B/32","Protocol":"finetune-proj","Input Modalities":"Image vs Caption ⊕ Definition"}. This record does not assert independent reproduction.

### gMLP

Idioms Accuracy: 33.7. Result ID: 6a217dbffbc373592c1efea1.

[Reporting paper](<https://www.sota2.com/research/paper/6a2106eec10484bac23f0669.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a101e209131cea25ed6.md>)

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

### RePercENT

Idioms Accuracy: 33.6. Result ID: 6a217dbffbc373592c1efea6.

[Reporting paper](<https://www.sota2.com/research/paper/6a2106eec10484bac23f0669.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a2106eec10484bac23f0669.md>)

Recorded evaluation configuration: {"SE (Semantic Encodings)":"true","GSA (Group Slot Attention)":"true"}. This record does not assert independent reproduction.

### RePercENT

Idioms Accuracy: 32.6. Result ID: 6a217dbffbc373592c1efea5.

[Reporting paper](<https://www.sota2.com/research/paper/6a2106eec10484bac23f0669.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a2106eec10484bac23f0669.md>)

Recorded evaluation configuration: {"SE (Semantic Encodings)":"true","GSA (Group Slot Attention)":"false"}. This record does not assert independent reproduction.

### RePercENT

Idioms Accuracy: 30.2. Result ID: 6a217dbffbc373592c1efea4.

[Reporting paper](<https://www.sota2.com/research/paper/6a2106eec10484bac23f0669.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a2106eec10484bac23f0669.md>)

Recorded evaluation configuration: {"SE (Semantic Encodings)":"false","GSA (Group Slot Attention)":"true"}. This record does not assert independent reproduction.

### RePercENT

Idioms Accuracy: 28.2. Result ID: 6a217dbffbc373592c1efea3.

[Reporting paper](<https://www.sota2.com/research/paper/6a2106eec10484bac23f0669.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a2106eec10484bac23f0669.md>)

Recorded evaluation configuration: {"SE (Semantic Encodings)":"false","GSA (Group Slot Attention)":"false"}. This record does not assert independent reproduction.

### CLIP

Idioms Accuracy: 26. Result ID: 6a217dbffbc373592c1efe90.

[Reporting paper](<https://www.sota2.com/research/paper/6a2106eec10484bac23f0669.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69581e209131cea21ad0.md>)

[Implementation](<https://github.com/OpenAI/CLIP>)

Recorded evaluation configuration: {"Backbone":"ViT-B/32","Protocol":"zero-shot","Input Modalities":"Image vs Caption ⊕ Definition"}. This record does not assert independent reproduction.

### CLIP-ViT-B/32

Idioms Accuracy: 22.5. Result ID: 6a217dbffbc373592c1efea0.

[Reporting paper](<https://www.sota2.com/research/paper/6a2106eec10484bac23f0669.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69581e209131cea21ad0.md>)

[Implementation](<https://github.com/OpenAI/CLIP>)

Recorded evaluation configuration: {"Evaluation Protocol":"end-to-end FT"}. This record does not assert independent reproduction.

### CLIP-ViT-B/32

Idioms Accuracy: 20.4. Result ID: 6a217dbffbc373592c1efe9f.

[Reporting paper](<https://www.sota2.com/research/paper/6a2106eec10484bac23f0669.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69581e209131cea21ad0.md>)

[Implementation](<https://github.com/OpenAI/CLIP>)

Recorded evaluation configuration: {"Evaluation Protocol":"projection-only FT"}. This record does not assert independent reproduction.

### CLIP-ViT-B/32

Idioms Accuracy: 16. Result ID: 6a217dbffbc373592c1efe9e.

[Reporting paper](<https://www.sota2.com/research/paper/6a2106eec10484bac23f0669.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69581e209131cea21ad0.md>)

[Implementation](<https://github.com/OpenAI/CLIP>)

Recorded evaluation configuration: {"Evaluation Protocol":"zero-shot"}. This record does not assert independent reproduction.

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