# Gendered-term mention ratio control on InclusiveGender

Source: [Gendered-term mention ratio control on InclusiveGender](<https://www.sota2.com/research/sota/gendered-term-mention-ratio-control-on-inclusivegender>)

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

Data updated: 2026-06-01T08:04:25.429Z

Evaluation of model performance on controlling the ratio of gendered-term mentions within the InclusiveGender dataset, measured by changes in mention ratios (Δ M, Δ F, Δ N).

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 | Gendered-term mention ratio control |
| Dataset (including any stated split) | InclusiveGender |
| Metric | Δ M Mention Ratio |
| Unit | SCALAR |
| Direction | Lower is better |
| Results loaded | 18 of 18 |
| 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

- [Δ M Mention Ratio](<https://www.sota2.com/research/sota/gendered-term-mention-ratio-control-on-inclusivegender.md?metric=%CE%94%20M%20Mention%20Ratio>): Lower is better; unit: SCALAR.

- [Δ F Mention Ratio](<https://www.sota2.com/research/sota/gendered-term-mention-ratio-control-on-inclusivegender.md?metric=%CE%94%20F%20Mention%20Ratio>): Lower is better; unit: SCALAR.

- [Δ N Mention Ratio](<https://www.sota2.com/research/sota/gendered-term-mention-ratio-control-on-inclusivegender.md?metric=%CE%94%20N%20Mention%20Ratio>): Lower is better; unit: SCALAR.

## Recorded results

| Method | Δ M Mention Ratio | Result date | Report paper |
| --- | --- | --- | --- |
| sNeuron-TST | -2.37 | Not recorded | 6a1d1213dbd7150506f83fef |
| LAPE | -2.03 | Not recorded | 6a1d1213dbd7150506f83fef |
| LAPE | -1.65 | Not recorded | 6a1d1213dbd7150506f83fef |
| LAPE | -0.36 | Not recorded | 6a1d1213dbd7150506f83fef |
| neuron-level intervention approach | -0.33 | Not recorded | 6a1d1213dbd7150506f83fef |
| LAPE | -0.16 | Not recorded | 6a1d1213dbd7150506f83fef |
| sNeuron-TST | -0.09 | Not recorded | 6a1d1213dbd7150506f83fef |
| neuron-level intervention approach | -0.04 | Not recorded | 6a1d1213dbd7150506f83fef |
| neuron-level intervention approach | -0.03 | Not recorded | 6a1d1213dbd7150506f83fef |
| neuron-level intervention approach | -0.02 | Not recorded | 6a1d1213dbd7150506f83fef |
| neuron-level intervention approach | 0.02 | Not recorded | 6a1d1213dbd7150506f83fef |
| LAPE | 0.02 | Not recorded | 6a1d1213dbd7150506f83fef |
| LAPE | 0.07 | Not recorded | 6a1d1213dbd7150506f83fef |
| sNeuron-TST | 0.17 | Not recorded | 6a1d1213dbd7150506f83fef |
| sNeuron-TST | 0.17 | Not recorded | 6a1d1213dbd7150506f83fef |
| neuron-level intervention approach | 0.33 | Not recorded | 6a1d1213dbd7150506f83fef |
| sNeuron-TST | 1.11 | Not recorded | 6a1d1213dbd7150506f83fef |
| sNeuron-TST | 3.54 | Not recorded | 6a1d1213dbd7150506f83fef |

### sNeuron-TST

Δ M Mention Ratio: -2.37. Result ID: 6a1d3c4cfbc373592c1639ee.

[Reporting paper](<https://www.sota2.com/research/paper/6a1d1213dbd7150506f83fef.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75081e209131cea65682.md>)

Recorded evaluation configuration: {"Model":"Llama 3.1","Target":"Masculine"}. This record does not assert independent reproduction.

### LAPE

Δ M Mention Ratio: -2.03. Result ID: 6a1d3c4cfbc373592c1639e6.

[Reporting paper](<https://www.sota2.com/research/paper/6a1d1213dbd7150506f83fef.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa738b1e209131cea5d1c2.md>)

[Implementation](<https://github.com/RUCAIBox/Language-Specific-Neurons>)

Recorded evaluation configuration: {"Model":"Qwen 2.5","Target":"Masculine"}. This record does not assert independent reproduction.

### LAPE

Δ M Mention Ratio: -1.65. Result ID: 6a1d3c4cfbc373592c1639ef.

[Reporting paper](<https://www.sota2.com/research/paper/6a1d1213dbd7150506f83fef.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa738b1e209131cea5d1c2.md>)

[Implementation](<https://github.com/RUCAIBox/Language-Specific-Neurons>)

Recorded evaluation configuration: {"Model":"Llama 3.1","Target":"Masculine"}. This record does not assert independent reproduction.

### LAPE

Δ M Mention Ratio: -0.36. Result ID: 6a1d3c4cfbc373592c1639ec.

[Reporting paper](<https://www.sota2.com/research/paper/6a1d1213dbd7150506f83fef.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa738b1e209131cea5d1c2.md>)

[Implementation](<https://github.com/RUCAIBox/Language-Specific-Neurons>)

Recorded evaluation configuration: {"Model":"Llama 3.1","Target":"Feminine"}. This record does not assert independent reproduction.

### neuron-level intervention approach

Δ M Mention Ratio: -0.33. Result ID: 6a1d3c4cfbc373592c1639ed.

[Reporting paper](<https://www.sota2.com/research/paper/6a1d1213dbd7150506f83fef.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a1d1213dbd7150506f83fef.md>)

Recorded evaluation configuration: {"Model":"Llama 3.1","Target":"Feminine"}. This record does not assert independent reproduction.

### LAPE

Δ M Mention Ratio: -0.16. Result ID: 6a1d3c4cfbc373592c1639e9.

[Reporting paper](<https://www.sota2.com/research/paper/6a1d1213dbd7150506f83fef.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa738b1e209131cea5d1c2.md>)

[Implementation](<https://github.com/RUCAIBox/Language-Specific-Neurons>)

Recorded evaluation configuration: {"Model":"Qwen 2.5","Target":"Neutral"}. This record does not assert independent reproduction.

### sNeuron-TST

Δ M Mention Ratio: -0.09. Result ID: 6a1d3c4cfbc373592c1639e2.

[Reporting paper](<https://www.sota2.com/research/paper/6a1d1213dbd7150506f83fef.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75081e209131cea65682.md>)

Recorded evaluation configuration: {"Model":"Qwen 2.5","Target":"Feminine"}. This record does not assert independent reproduction.

### neuron-level intervention approach

Δ M Mention Ratio: -0.04. Result ID: 6a1d3c4cfbc373592c1639e4.

[Reporting paper](<https://www.sota2.com/research/paper/6a1d1213dbd7150506f83fef.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a1d1213dbd7150506f83fef.md>)

Recorded evaluation configuration: {"Model":"Qwen 2.5","Target":"Feminine"}. This record does not assert independent reproduction.

### neuron-level intervention approach

Δ M Mention Ratio: -0.03. Result ID: 6a1d3c4cfbc373592c1639ea.

[Reporting paper](<https://www.sota2.com/research/paper/6a1d1213dbd7150506f83fef.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a1d1213dbd7150506f83fef.md>)

Recorded evaluation configuration: {"Model":"Qwen 2.5","Target":"Neutral"}. This record does not assert independent reproduction.

### neuron-level intervention approach

Δ M Mention Ratio: -0.02. Result ID: 6a1d3c4cfbc373592c1639f3.

[Reporting paper](<https://www.sota2.com/research/paper/6a1d1213dbd7150506f83fef.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a1d1213dbd7150506f83fef.md>)

Recorded evaluation configuration: {"Model":"Llama 3.1","Target":"Neutral"}. This record does not assert independent reproduction.

### neuron-level intervention approach

Δ M Mention Ratio: 0.02. Result ID: 6a1d3c4cfbc373592c1639e7.

[Reporting paper](<https://www.sota2.com/research/paper/6a1d1213dbd7150506f83fef.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a1d1213dbd7150506f83fef.md>)

Recorded evaluation configuration: {"Model":"Qwen 2.5","Target":"Masculine"}. This record does not assert independent reproduction.

### LAPE

Δ M Mention Ratio: 0.02. Result ID: 6a1d3c4cfbc373592c1639f2.

[Reporting paper](<https://www.sota2.com/research/paper/6a1d1213dbd7150506f83fef.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa738b1e209131cea5d1c2.md>)

[Implementation](<https://github.com/RUCAIBox/Language-Specific-Neurons>)

Recorded evaluation configuration: {"Model":"Llama 3.1","Target":"Neutral"}. This record does not assert independent reproduction.

### LAPE

Δ M Mention Ratio: 0.07. Result ID: 6a1d3c4cfbc373592c1639e3.

[Reporting paper](<https://www.sota2.com/research/paper/6a1d1213dbd7150506f83fef.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa738b1e209131cea5d1c2.md>)

[Implementation](<https://github.com/RUCAIBox/Language-Specific-Neurons>)

Recorded evaluation configuration: {"Model":"Qwen 2.5","Target":"Feminine"}. This record does not assert independent reproduction.

### sNeuron-TST

Δ M Mention Ratio: 0.17. Result ID: 6a1d3c4cfbc373592c1639e5.

[Reporting paper](<https://www.sota2.com/research/paper/6a1d1213dbd7150506f83fef.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75081e209131cea65682.md>)

Recorded evaluation configuration: {"Model":"Qwen 2.5","Target":"Masculine"}. This record does not assert independent reproduction.

### sNeuron-TST

Δ M Mention Ratio: 0.17. Result ID: 6a1d3c4cfbc373592c1639f1.

[Reporting paper](<https://www.sota2.com/research/paper/6a1d1213dbd7150506f83fef.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75081e209131cea65682.md>)

Recorded evaluation configuration: {"Model":"Llama 3.1","Target":"Neutral"}. This record does not assert independent reproduction.

### neuron-level intervention approach

Δ M Mention Ratio: 0.33. Result ID: 6a1d3c4cfbc373592c1639f0.

[Reporting paper](<https://www.sota2.com/research/paper/6a1d1213dbd7150506f83fef.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a1d1213dbd7150506f83fef.md>)

Recorded evaluation configuration: {"Model":"Llama 3.1","Target":"Masculine"}. This record does not assert independent reproduction.

### sNeuron-TST

Δ M Mention Ratio: 1.11. Result ID: 6a1d3c4cfbc373592c1639e8.

[Reporting paper](<https://www.sota2.com/research/paper/6a1d1213dbd7150506f83fef.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75081e209131cea65682.md>)

Recorded evaluation configuration: {"Model":"Qwen 2.5","Target":"Neutral"}. This record does not assert independent reproduction.

### sNeuron-TST

Δ M Mention Ratio: 3.54. Result ID: 6a1d3c4cfbc373592c1639eb.

[Reporting paper](<https://www.sota2.com/research/paper/6a1d1213dbd7150506f83fef.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75081e209131cea65682.md>)

Recorded evaluation configuration: {"Model":"Llama 3.1","Target":"Feminine"}. This record does not assert independent reproduction.

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