# Legal Reasoning on LegalArg

Source: [Legal Reasoning on LegalArg](<https://www.sota2.com/research/sota/legal-reasoning-on-legalarg>)

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

Data updated: 2026-03-19T09:16:41.548Z

Evaluation of model performance on Legal Reasoning tasks using the LegalArg dataset, 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 | Legal Reasoning |
| Dataset (including any stated split) | LegalArg |
| Metric | Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 14 of 14 |
| 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/legal-reasoning-on-legalarg.md?metric=Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| Qwen3-4B-Instruct | 65.42 | Not recorded | 69bb83200636e5bf4afc1b2d |
| Qwen3-8B-Arguinas-Target-SFT | 62.87 | Not recorded | 69bb83200636e5bf4afc1b2d |
| Qwen3-4B-Instruct-Arguinas-SFT | 61.75 | Not recorded | 69bb83200636e5bf4afc1b2d |
| Qwen2.5-7B-Instruct-Arguinas-SFT | 61.67 | Not recorded | 69bb83200636e5bf4afc1b2d |
| Qwen3-8B-AAAC-Target-SFT | 61.63 | Not recorded | 69bb83200636e5bf4afc1b2d |
| Qwen3-8B-EntailmentBank-Target-SFT | 61.24 | Not recorded | 69bb83200636e5bf4afc1b2d |
| Qwen2.5-7B-Instruct | 60.19 | Not recorded | 69bb83200636e5bf4afc1b2d |
| Qwen3-4B-Arguinas-Target-SFT | 58.52 | Not recorded | 69bb83200636e5bf4afc1b2d |
| Qwen3-8B-ArgumentOnly-Target-SFT | 57.89 | Not recorded | 69bb83200636e5bf4afc1b2d |
| Qwen3-4B-AAAC-Target-SFT | 56.06 | Not recorded | 69bb83200636e5bf4afc1b2d |
| Qwen3-4B-EntailmentBank-Target-SFT | 55.53 | Not recorded | 69bb83200636e5bf4afc1b2d |
| Qwen3-8B-Target-SFT | 54.7 | Not recorded | 69bb83200636e5bf4afc1b2d |
| Qwen3-4B-ArgumentOnly-Target-SFT | 48.6 | Not recorded | 69bb83200636e5bf4afc1b2d |
| Qwen3-4B-Target-SFT | 47.24 | Not recorded | 69bb83200636e5bf4afc1b2d |

### Qwen3-4B-Instruct

Accuracy: 65.42. Result ID: 69bbbef447f34ce924a5e18b.

[Reporting paper](<https://www.sota2.com/research/paper/69bb83200636e5bf4afc1b2d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa788e1e209131cea79116.md>)

Recorded evaluation configuration: {"Model Backbone":"Qwen3-4B-Instruct","Finetuning Strategy":"Continued Finetuning","SFT Data Source":"Baseline"}. This record does not assert independent reproduction.

### Qwen3-8B-Arguinas-Target-SFT

Accuracy: 62.87. Result ID: 69bbbef447f34ce924a5e18a.

[Reporting paper](<https://www.sota2.com/research/paper/69bb83200636e5bf4afc1b2d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69bb83200636e5bf4afc1b2d.md>)

Recorded evaluation configuration: {"Model Backbone":"Qwen3-8B","Finetuning Strategy":"Pre-adaptive Finetuning","SFT Data Source":"Arguinas"}. This record does not assert independent reproduction.

### Qwen3-4B-Instruct-Arguinas-SFT

Accuracy: 61.75. Result ID: 69bbbef447f34ce924a5e18c.

[Reporting paper](<https://www.sota2.com/research/paper/69bb83200636e5bf4afc1b2d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69bb83200636e5bf4afc1b2d.md>)

Recorded evaluation configuration: {"Model Backbone":"Qwen3-4B-Instruct","Finetuning Strategy":"Continued Finetuning","SFT Data Source":"Arguinas"}. This record does not assert independent reproduction.

### Qwen2.5-7B-Instruct-Arguinas-SFT

Accuracy: 61.67. Result ID: 69bbbef447f34ce924a5e18e.

[Reporting paper](<https://www.sota2.com/research/paper/69bb83200636e5bf4afc1b2d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69bb83200636e5bf4afc1b2d.md>)

Recorded evaluation configuration: {"Model Backbone":"Qwen2.5-7B-Instruct","Finetuning Strategy":"Continued Finetuning","SFT Data Source":"Arguinas"}. This record does not assert independent reproduction.

### Qwen3-8B-AAAC-Target-SFT

Accuracy: 61.63. Result ID: 69bbbef447f34ce924a5e188.

[Reporting paper](<https://www.sota2.com/research/paper/69bb83200636e5bf4afc1b2d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa788e1e209131cea79116.md>)

Recorded evaluation configuration: {"Model Backbone":"Qwen3-8B","Finetuning Strategy":"Pre-adaptive Finetuning","SFT Data Source":"AAAC"}. This record does not assert independent reproduction.

### Qwen3-8B-EntailmentBank-Target-SFT

Accuracy: 61.24. Result ID: 69bbbef447f34ce924a5e189.

[Reporting paper](<https://www.sota2.com/research/paper/69bb83200636e5bf4afc1b2d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa788e1e209131cea79116.md>)

Recorded evaluation configuration: {"Model Backbone":"Qwen3-8B","Finetuning Strategy":"Pre-adaptive Finetuning","SFT Data Source":"EntailmentBank"}. This record does not assert independent reproduction.

### Qwen2.5-7B-Instruct

Accuracy: 60.19. Result ID: 69bbbef447f34ce924a5e18d.

[Reporting paper](<https://www.sota2.com/research/paper/69bb83200636e5bf4afc1b2d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa767e1e209131cea6d6a7.md>)

Recorded evaluation configuration: {"Model Backbone":"Qwen2.5-7B-Instruct","Finetuning Strategy":"Continued Finetuning","SFT Data Source":"Baseline"}. This record does not assert independent reproduction.

### Qwen3-4B-Arguinas-Target-SFT

Accuracy: 58.52. Result ID: 69bbbef447f34ce924a5e185.

[Reporting paper](<https://www.sota2.com/research/paper/69bb83200636e5bf4afc1b2d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69bb83200636e5bf4afc1b2d.md>)

Recorded evaluation configuration: {"Model Backbone":"Qwen3-4B","Finetuning Strategy":"Pre-adaptive Finetuning","SFT Data Source":"Arguinas"}. This record does not assert independent reproduction.

### Qwen3-8B-ArgumentOnly-Target-SFT

Accuracy: 57.89. Result ID: 69bbbef447f34ce924a5e187.

[Reporting paper](<https://www.sota2.com/research/paper/69bb83200636e5bf4afc1b2d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa788e1e209131cea79116.md>)

Recorded evaluation configuration: {"Model Backbone":"Qwen3-8B","Finetuning Strategy":"Pre-adaptive Finetuning","SFT Data Source":"ArgumentOnly"}. This record does not assert independent reproduction.

### Qwen3-4B-AAAC-Target-SFT

Accuracy: 56.06. Result ID: 69bbbef447f34ce924a5e183.

[Reporting paper](<https://www.sota2.com/research/paper/69bb83200636e5bf4afc1b2d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa788e1e209131cea79116.md>)

Recorded evaluation configuration: {"Model Backbone":"Qwen3-4B","Finetuning Strategy":"Pre-adaptive Finetuning","SFT Data Source":"AAAC"}. This record does not assert independent reproduction.

### Qwen3-4B-EntailmentBank-Target-SFT

Accuracy: 55.53. Result ID: 69bbbef447f34ce924a5e184.

[Reporting paper](<https://www.sota2.com/research/paper/69bb83200636e5bf4afc1b2d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa788e1e209131cea79116.md>)

Recorded evaluation configuration: {"Model Backbone":"Qwen3-4B","Finetuning Strategy":"Pre-adaptive Finetuning","SFT Data Source":"EntailmentBank"}. This record does not assert independent reproduction.

### Qwen3-8B-Target-SFT

Accuracy: 54.7. Result ID: 69bbbef447f34ce924a5e186.

[Reporting paper](<https://www.sota2.com/research/paper/69bb83200636e5bf4afc1b2d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa788e1e209131cea79116.md>)

Recorded evaluation configuration: {"Model Backbone":"Qwen3-8B","Finetuning Strategy":"Pre-adaptive Finetuning","SFT Data Source":"Target"}. This record does not assert independent reproduction.

### Qwen3-4B-ArgumentOnly-Target-SFT

Accuracy: 48.6. Result ID: 69bbbef447f34ce924a5e182.

[Reporting paper](<https://www.sota2.com/research/paper/69bb83200636e5bf4afc1b2d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa788e1e209131cea79116.md>)

Recorded evaluation configuration: {"Model Backbone":"Qwen3-4B","Finetuning Strategy":"Pre-adaptive Finetuning","SFT Data Source":"ArgumentOnly"}. This record does not assert independent reproduction.

### Qwen3-4B-Target-SFT

Accuracy: 47.24. Result ID: 69bbbef447f34ce924a5e181.

[Reporting paper](<https://www.sota2.com/research/paper/69bb83200636e5bf4afc1b2d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa788e1e209131cea79116.md>)

Recorded evaluation configuration: {"Model Backbone":"Qwen3-4B","Finetuning Strategy":"Pre-adaptive Finetuning","SFT Data Source":"Target"}. This record does not assert independent reproduction.

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