# Factuality Evaluation on AggreFact (FTSOTA)

Source: [Factuality Evaluation on AggreFact (FTSOTA)](<https://www.sota2.com/research/sota/factuality-evaluation-on-aggrefact-ftsota>)

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

Data updated: 2026-02-26T05:37:36.365Z

Evaluation of model performance on the AggreFact (FTSOTA) dataset for factuality assessment, typically measured using Balanced Accuracy across CNN-FTS, XSUM-FTS, and average 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 | Factuality Evaluation |
| Dataset (including any stated split) | AggreFact (FTSOTA) |
| Metric | Balanced Accuracy (CNN-FTS) |
| 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

- [Balanced Accuracy (CNN-FTS)](<https://www.sota2.com/research/sota/factuality-evaluation-on-aggrefact-ftsota.md?metric=Balanced%20Accuracy%20(CNN-FTS)>): Higher is better; unit: PERCENT.

- [Balanced Accuracy (XSUM-FTS)](<https://www.sota2.com/research/sota/factuality-evaluation-on-aggrefact-ftsota.md?metric=Balanced%20Accuracy%20(XSUM-FTS)>): Higher is better; unit: PERCENT.

- [Balanced Accuracy (AVG)](<https://www.sota2.com/research/sota/factuality-evaluation-on-aggrefact-ftsota.md?metric=Balanced%20Accuracy%20(AVG)>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Balanced Accuracy (CNN-FTS) | Result date | Report paper |
| --- | --- | --- | --- |
| FENICE\_GPT\_claims | 70.5 | Not recorded | 68fa74ae1e209131cea632b2 |
| QuestEval | 70.2 | Not recorded | 68fa74ae1e209131cea632b2 |
| QAFactEval | 67.8 | Not recorded | 68fa74ae1e209131cea632b2 |
| FENICE\_T5\_claims | 67.7 | Not recorded | 68fa74ae1e209131cea632b2 |
| DAE | 65.4 | Not recorded | 68fa74ae1e209131cea632b2 |
| SummaC-ZS | 64 | Not recorded | 68fa74ae1e209131cea632b2 |
| MENLI | 63.4 | Not recorded | 68fa74ae1e209131cea632b2 |
| AlignScore | 62.7 | Not recorded | 68fa74ae1e209131cea632b2 |
| TrueTeacher-11B | 62 | Not recorded | 68fa74ae1e209131cea632b2 |
| SummaC-Conv | 61 | Not recorded | 68fa74ae1e209131cea632b2 |
| ChatGPT-ZS | 56.3 | Not recorded | 68fa74ae1e209131cea632b2 |
| ChatGPT-Star | 56.3 | Not recorded | 68fa74ae1e209131cea632b2 |
| ChatGPT-DA | 53.7 | Not recorded | 68fa74ae1e209131cea632b2 |
| ChatGPT-COT | 52.5 | Not recorded | 68fa74ae1e209131cea632b2 |

### FENICE\_GPT\_claims

Balanced Accuracy (CNN-FTS): 70.5. Result ID: 6993540b3a6c314849153772.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74ae1e209131cea632b2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74ae1e209131cea632b2.md>)

[Implementation](<https://huggingface.co/MoritzLaurer/DeBERTa-v3-large-mnli-fever-anli-ling-wanli>)

Recorded evaluation configuration: {"thresholding":"single-threshold","claim\_extractor":"LLM-based (GPT)"}. This record does not assert independent reproduction.

### QuestEval

Balanced Accuracy (CNN-FTS): 70.2. Result ID: 6993540b3a6c314849153767.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74ae1e209131cea632b2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69b91e209131cea23bee.md>)

Recorded evaluation configuration: {"thresholding":"single-threshold"}. This record does not assert independent reproduction.

### QAFactEval

Balanced Accuracy (CNN-FTS): 67.8. Result ID: 6993540b3a6c31484915376a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74ae1e209131cea632b2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c611e209131cea33b79.md>)

Recorded evaluation configuration: {"thresholding":"single-threshold"}. This record does not assert independent reproduction.

### FENICE\_T5\_claims

Balanced Accuracy (CNN-FTS): 67.7. Result ID: 6993540b3a6c314849153773.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74ae1e209131cea632b2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74ae1e209131cea632b2.md>)

[Implementation](<https://huggingface.co/MoritzLaurer/DeBERTa-v3-large-mnli-fever-anli-ling-wanli>)

Recorded evaluation configuration: {"thresholding":"single-threshold","claim\_extractor":"knowledge-distilled (T5)"}. This record does not assert independent reproduction.

### DAE

Balanced Accuracy (CNN-FTS): 65.4. Result ID: 6993540b3a6c314849153766.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74ae1e209131cea632b2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa68891e209131cea1cc4b.md>)

Recorded evaluation configuration: {"thresholding":"single-threshold"}. This record does not assert independent reproduction.

### SummaC-ZS

Balanced Accuracy (CNN-FTS): 64. Result ID: 6993540b3a6c314849153768.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74ae1e209131cea632b2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b421e209131cea2d015.md>)

Recorded evaluation configuration: {"thresholding":"single-threshold"}. This record does not assert independent reproduction.

### MENLI

Balanced Accuracy (CNN-FTS): 63.4. Result ID: 6993540b3a6c31484915376c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74ae1e209131cea632b2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71a91e209131cea52674.md>)

Recorded evaluation configuration: {"thresholding":"single-threshold"}. This record does not assert independent reproduction.

### AlignScore

Balanced Accuracy (CNN-FTS): 62.7. Result ID: 6993540b3a6c31484915376d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74ae1e209131cea632b2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f801e209131cea4607f.md>)

[Implementation](<https://github.com/yuh-zha/AlignScore>)

Recorded evaluation configuration: {"thresholding":"single-threshold"}. This record does not assert independent reproduction.

### TrueTeacher-11B

Balanced Accuracy (CNN-FTS): 62. Result ID: 6993540b3a6c31484915376b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74ae1e209131cea632b2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70e61e209131cea4dd06.md>)

Recorded evaluation configuration: {"thresholding":"single-threshold","parameters":"11B"}. This record does not assert independent reproduction.

### SummaC-Conv

Balanced Accuracy (CNN-FTS): 61. Result ID: 6993540b3a6c314849153769.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74ae1e209131cea632b2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b421e209131cea2d015.md>)

Recorded evaluation configuration: {"thresholding":"single-threshold"}. This record does not assert independent reproduction.

### ChatGPT-ZS

Balanced Accuracy (CNN-FTS): 56.3. Result ID: 6993540b3a6c31484915376e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74ae1e209131cea632b2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f061e209131cea4358f.md>)

Recorded evaluation configuration: {"thresholding":"single-threshold","mode":"zero-shot"}. This record does not assert independent reproduction.

### ChatGPT-Star

Balanced Accuracy (CNN-FTS): 56.3. Result ID: 6993540b3a6c314849153771.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74ae1e209131cea632b2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70f71e209131cea4e706.md>)

Recorded evaluation configuration: {"thresholding":"single-threshold"}. This record does not assert independent reproduction.

### ChatGPT-DA

Balanced Accuracy (CNN-FTS): 53.7. Result ID: 6993540b3a6c314849153770.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74ae1e209131cea632b2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70f71e209131cea4e706.md>)

Recorded evaluation configuration: {"thresholding":"single-threshold"}. This record does not assert independent reproduction.

### ChatGPT-COT

Balanced Accuracy (CNN-FTS): 52.5. Result ID: 6993540b3a6c31484915376f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74ae1e209131cea632b2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f061e209131cea4358f.md>)

Recorded evaluation configuration: {"thresholding":"single-threshold","mode":"chain-of-thought"}. This record does not assert independent reproduction.

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