# Faithfulness Measurement on tldr\_news

Source: [Faithfulness Measurement on tldr\_news](<https://www.sota2.com/research/sota/faithfulness-measurement-on-tldr_news>)

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

Data updated: 2026-02-26T05:36:45.565Z

Evaluation of model faithfulness on the tldr\_news dataset, measured using metrics such as BLEU, ROUGE-L, SentenceBert, PR, and KL divergence.

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 | Faithfulness Measurement |
| Dataset (including any stated split) | tldr\_news |
| Metric | BLEU |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 12 of 12 |
| 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

- [BLEU](<https://www.sota2.com/research/sota/faithfulness-measurement-on-tldr_news.md?metric=BLEU>): Higher is better; unit: PERCENT.

- [ROUGE-L Precision](<https://www.sota2.com/research/sota/faithfulness-measurement-on-tldr_news.md?metric=ROUGE-L%20Precision>): Higher is better; unit: PERCENT.

- [ROUGE-L Recall](<https://www.sota2.com/research/sota/faithfulness-measurement-on-tldr_news.md?metric=ROUGE-L%20Recall>): Higher is better; unit: PERCENT.

- [ROUGE-L F1](<https://www.sota2.com/research/sota/faithfulness-measurement-on-tldr_news.md?metric=ROUGE-L%20F1>): Higher is better; unit: PERCENT.

- [SentenceBert Similarity](<https://www.sota2.com/research/sota/faithfulness-measurement-on-tldr_news.md?metric=SentenceBert%20Similarity>): Higher is better; unit: PERCENT.

- [PR](<https://www.sota2.com/research/sota/faithfulness-measurement-on-tldr_news.md?metric=PR>): Higher is better; unit: PERCENT.

- [KL Divergence](<https://www.sota2.com/research/sota/faithfulness-measurement-on-tldr_news.md?metric=KL%20Divergence>): Lower is better; unit: SCALAR.

## Recorded results

| Method | BLEU | Result date | Report paper |
| --- | --- | --- | --- |
| Random | 79.4 | Not recorded | 68fa7b651e209131cea882f8 |
| Random | 78.1 | Not recorded | 68fa7b651e209131cea882f8 |
| Attention | 76.7 | Not recorded | 68fa7b651e209131cea882f8 |
| Captum | 75.9 | Not recorded | 68fa7b651e209131cea882f8 |
| Last-Attention | 74.7 | Not recorded | 68fa7b651e209131cea882f8 |
| Last-Attention | 74.6 | Not recorded | 68fa7b651e209131cea882f8 |
| Integrated-Gradient | 71.3 | Not recorded | 68fa7b651e209131cea882f8 |
| Integrated-Gradient | 70.5 | Not recorded | 68fa7b651e209131cea882f8 |
| JoPA | 69.2 | Not recorded | 68fa7b651e209131cea882f8 |
| Attention | 62.1 | Not recorded | 68fa7b651e209131cea882f8 |
| Captum | 55.6 | Not recorded | 68fa7b651e209131cea882f8 |
| JoPA | 53.6 | Not recorded | 68fa7b651e209131cea882f8 |

### Random

BLEU: 79.4. Result ID: 698b852e7a967eddb76f42d6.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7b651e209131cea882f8.md>)

Recorded evaluation configuration: {"Model":"LlaMA-2 (7B-Chat)"}. This record does not assert independent reproduction.

### Random

BLEU: 78.1. Result ID: 698b852e7a967eddb76f42dc.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7b651e209131cea882f8.md>)

Recorded evaluation configuration: {"Model":"Vicuna (7B)"}. This record does not assert independent reproduction.

### Attention

BLEU: 76.7. Result ID: 698b852e7a967eddb76f42d8.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7b651e209131cea882f8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67511e209131cea15642.md>)

Recorded evaluation configuration: {"Model":"LlaMA-2 (7B-Chat)"}. This record does not assert independent reproduction.

### Captum

BLEU: 75.9. Result ID: 698b852e7a967eddb76f42d9.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7b651e209131cea882f8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71611e209131cea51003.md>)

Recorded evaluation configuration: {"Model":"LlaMA-2 (7B-Chat)"}. This record does not assert independent reproduction.

### Last-Attention

BLEU: 74.7. Result ID: 698b852e7a967eddb76f42d7.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7b651e209131cea882f8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72231e209131cea54be7.md>)

Recorded evaluation configuration: {"Model":"LlaMA-2 (7B-Chat)"}. This record does not assert independent reproduction.

### Last-Attention

BLEU: 74.6. Result ID: 698b852e7a967eddb76f42dd.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7b651e209131cea882f8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72231e209131cea54be7.md>)

Recorded evaluation configuration: {"Model":"Vicuna (7B)"}. This record does not assert independent reproduction.

### Integrated-Gradient

BLEU: 71.3. Result ID: 698b852e7a967eddb76f42da.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7b651e209131cea882f8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa63661e209131ce9fe6bd.md>)

Recorded evaluation configuration: {"Model":"LlaMA-2 (7B-Chat)"}. This record does not assert independent reproduction.

### Integrated-Gradient

BLEU: 70.5. Result ID: 698b852e7a967eddb76f42e0.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7b651e209131cea882f8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa63661e209131ce9fe6bd.md>)

Recorded evaluation configuration: {"Model":"Vicuna (7B)"}. This record does not assert independent reproduction.

### JoPA

BLEU: 69.2. Result ID: 698b852e7a967eddb76f42db.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7b651e209131cea882f8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7b651e209131cea882f8.md>)

[Implementation](<https://github.com/yuruic/JoPA>)

Recorded evaluation configuration: {"Model":"LlaMA-2 (7B-Chat)"}. This record does not assert independent reproduction.

### Attention

BLEU: 62.1. Result ID: 698b852e7a967eddb76f42de.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7b651e209131cea882f8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67511e209131cea15642.md>)

Recorded evaluation configuration: {"Model":"Vicuna (7B)"}. This record does not assert independent reproduction.

### Captum

BLEU: 55.6. Result ID: 698b852e7a967eddb76f42df.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7b651e209131cea882f8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71611e209131cea51003.md>)

Recorded evaluation configuration: {"Model":"Vicuna (7B)"}. This record does not assert independent reproduction.

### JoPA

BLEU: 53.6. Result ID: 698b852e7a967eddb76f42e1.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7b651e209131cea882f8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7b651e209131cea882f8.md>)

[Implementation](<https://github.com/yuruic/JoPA>)

Recorded evaluation configuration: {"Model":"Vicuna (7B)"}. This record does not assert independent reproduction.

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