# Multi-aspect Controllable Text Generation on CompMCTG 1.0 (Hold-Out)

Source: [Multi-aspect Controllable Text Generation on CompMCTG 1.0 (Hold-Out)](<https://www.sota2.com/research/sota/multi-aspect-controllable-text-generation-on-compmctg-1-0-hold-out>)

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

Data updated: 2026-02-26T05:34:46.917Z

Evaluation of Multi-aspect Controllable Text Generation performance on the CompMCTG 1.0 (Hold-Out) split, measured using Aid., Pid., Acomp, and Pcomp scores.

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 | Multi-aspect Controllable Text Generation |
| Dataset (including any stated split) | CompMCTG 1.0 (Hold-Out) |
| Metric | Aid Score |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 20 of 20 |
| 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

- [Aid Score](<https://www.sota2.com/research/sota/multi-aspect-controllable-text-generation-on-compmctg-1-0-hold-out.md?metric=Aid%20Score>): Higher is better; unit: PERCENT.

- [Pid Score](<https://www.sota2.com/research/sota/multi-aspect-controllable-text-generation-on-compmctg-1-0-hold-out.md?metric=Pid%20Score>): Higher is better; unit: PERCENT.

- [Acomp Score](<https://www.sota2.com/research/sota/multi-aspect-controllable-text-generation-on-compmctg-1-0-hold-out.md?metric=Acomp%20Score>): Higher is better; unit: PERCENT.

- [Pcomp Score](<https://www.sota2.com/research/sota/multi-aspect-controllable-text-generation-on-compmctg-1-0-hold-out.md?metric=Pcomp%20Score>): Higher is better; unit: PERCENT.

- [Relevance (i.d.)](<https://www.sota2.com/research/sota/multi-aspect-controllable-text-generation-on-compmctg-1-0-hold-out.md?metric=Relevance%20(i.d.)>): Higher is better; unit: SCALAR.

- [Fluency (i.d.)](<https://www.sota2.com/research/sota/multi-aspect-controllable-text-generation-on-compmctg-1-0-hold-out.md?metric=Fluency%20(i.d.)>): Higher is better; unit: SCALAR.

- [Relevance (comp)](<https://www.sota2.com/research/sota/multi-aspect-controllable-text-generation-on-compmctg-1-0-hold-out.md?metric=Relevance%20(comp)>): Higher is better; unit: SCALAR.

- [Fluency (comp)](<https://www.sota2.com/research/sota/multi-aspect-controllable-text-generation-on-compmctg-1-0-hold-out.md?metric=Fluency%20(comp)>): Higher is better; unit: SCALAR.

## Recorded results

| Method | Aid Score | Result date | Report paper |
| --- | --- | --- | --- |
| Dis-Lens | 84.84 | Not recorded | 68fa73671e209131cea5c3a8 |
| Con. Prefix | 83.75 | Not recorded | 68fa73671e209131cea5c3a8 |
| DCG | 79.72 | Not recorded | 68fa73671e209131cea5c3a8 |
| CTRL | 78.89 | Not recorded | 68fa73671e209131cea5c3a8 |
| Prior | 73.64 | Not recorded | 68fa73671e209131cea5c3a8 |
| CatPrompt | 63.95 | Not recorded | 68fa73671e209131cea5c3a8 |
| LLAMA-2 | 62.61 | Not recorded | 68fa73671e209131cea5c3a8 |
| Fudge | 59.35 | Not recorded | 68fa73671e209131cea5c3a8 |
| ChatGPT | 56.62 | Not recorded | 68fa73671e209131cea5c3a8 |
| PPLM | 41.05 | Not recorded | 68fa73671e209131cea5c3a8 |

### Dis-Lens

Aid Score: 84.84. Result ID: 6992c7293a6c314849116ccd.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73671e209131cea5c3a8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6da91e209131cea3b43c.md>)

Recorded evaluation configuration: {"Approach Type":"Separate-Training based"}. This record does not assert independent reproduction.

### Con. Prefix

Aid Score: 83.75. Result ID: 6992c7293a6c314849116cd1.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73671e209131cea5c3a8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6bdd1e209131cea30afd.md>)

Recorded evaluation configuration: {"Approach Type":"Joint-Training based"}. This record does not assert independent reproduction.

### DCG

Aid Score: 79.72. Result ID: 6992c7293a6c314849116cd2.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73671e209131cea5c3a8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6fcd1e209131cea47828.md>)

[Implementation](<https://github.com/Zeng-WH/Seen-to-Unseen.>)

Recorded evaluation configuration: {"Approach Type":"Joint-Training based"}. This record does not assert independent reproduction.

### CTRL

Aid Score: 78.89. Result ID: 6992c7293a6c314849116ccf.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73671e209131cea5c3a8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa665f1e209131cea0fe3b.md>)

Recorded evaluation configuration: {"Approach Type":"Joint-Training based"}. This record does not assert independent reproduction.

### Prior

Aid Score: 73.64. Result ID: 6992c7293a6c314849116cce.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73671e209131cea5c3a8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f6b1e209131cea45727.md>)

[Implementation](<https://github.com/HappyGu0524/MultiControl>)

Recorded evaluation configuration: {"Approach Type":"Separate-Training based"}. This record does not assert independent reproduction.

### CatPrompt

Aid Score: 63.95. Result ID: 6992c7293a6c314849116cd0.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73671e209131cea5c3a8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c601e209131cea33ae1.md>)

Recorded evaluation configuration: {"Approach Type":"Joint-Training based"}. This record does not assert independent reproduction.

### LLAMA-2

Aid Score: 62.61. Result ID: 6992c7293a6c314849116cc9.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73671e209131cea5c3a8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa701f1e209131cea492fe.md>)

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

Recorded evaluation configuration: {"Approach Type":"LLM+In-context Learning"}. This record does not assert independent reproduction.

### Fudge

Aid Score: 59.35. Result ID: 6992c7293a6c314849116ccc.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73671e209131cea5c3a8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a931e209131cea29040.md>)

Recorded evaluation configuration: {"Approach Type":"Decoding-Time based"}. This record does not assert independent reproduction.

### ChatGPT

Aid Score: 56.62. Result ID: 6992c7293a6c314849116cca.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73671e209131cea5c3a8.md>)

Recorded evaluation configuration: {"Approach Type":"LLM+In-context Learning"}. This record does not assert independent reproduction.

### PPLM

Aid Score: 41.05. Result ID: 6992c7293a6c314849116ccb.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73671e209131cea5c3a8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67241e209131cea149df.md>)

Recorded evaluation configuration: {"Approach Type":"Decoding-Time based"}. This record does not assert independent reproduction.

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