# Faithfulness Measurement on MHC Source: [Faithfulness Measurement on MHC]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:36:46.703Z Evaluation of model faithfulness using the MHC dataset, measured by metrics including 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) | MHC | | Metric | BLEU | | Unit | PERCENT | | Direction | Higher 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 - [BLEU](): Higher is better; unit: PERCENT. - [ROUGE-L Precision](): Higher is better; unit: PERCENT. - [ROUGE-L Recall](): Higher is better; unit: PERCENT. - [ROUGE-L F1](): Higher is better; unit: PERCENT. - [SentenceBert Similarity](): Higher is better; unit: PERCENT. - [PR AUC](): Higher is better; unit: PERCENT. - [KL Divergence](): Lower is better; unit: SCALAR. ## Recorded results | Method | BLEU | Result date | Report paper | | --- | --- | --- | --- | | Random | 72.3 | Not recorded | 68fa7b651e209131cea882f8 | | Random | 71.5 | Not recorded | 68fa7b651e209131cea882f8 | | Random | 70.6 | Not recorded | 68fa7b651e209131cea882f8 | | Captum | 68.8 | Not recorded | 68fa7b651e209131cea882f8 | | Attention | 68.5 | Not recorded | 68fa7b651e209131cea882f8 | | Last-Attention | 68.4 | Not recorded | 68fa7b651e209131cea882f8 | | Integrated-Gradient | 67.2 | Not recorded | 68fa7b651e209131cea882f8 | | Attention | 66.5 | Not recorded | 68fa7b651e209131cea882f8 | | Attention | 66.4 | Not recorded | 68fa7b651e209131cea882f8 | | Last-Attention | 66.2 | Not recorded | 68fa7b651e209131cea882f8 | | Last-Attention | 66 | Not recorded | 68fa7b651e209131cea882f8 | | Integrated-Gradient | 64.6 | Not recorded | 68fa7b651e209131cea882f8 | | Captum | 64 | Not recorded | 68fa7b651e209131cea882f8 | | Integrated-Gradient | 61 | Not recorded | 68fa7b651e209131cea882f8 | | Captum | 57.9 | Not recorded | 68fa7b651e209131cea882f8 | | JoPA | 57.9 | Not recorded | 68fa7b651e209131cea882f8 | | JoPA | 57.5 | Not recorded | 68fa7b651e209131cea882f8 | | JoPA | 57.5 | Not recorded | 68fa7b651e209131cea882f8 | ### Random BLEU: 72.3. Result ID: 698b852e7a967eddb76f42e2. [Reporting paper]() Recorded evaluation configuration: {"Model":"LlaMA-2 (7B-Chat)"}. This record does not assert independent reproduction. ### Random BLEU: 71.5. Result ID: 698b852e7a967eddb76f42e8. [Reporting paper]() Recorded evaluation configuration: {"Model":"Vicuna (7B)"}. This record does not assert independent reproduction. ### Random BLEU: 70.6. Result ID: 698c5a2137894ef7d9cdbe18. [Reporting paper]() Recorded evaluation configuration: {"Model":"Llama-2 (7B-Chat)"}. This record does not assert independent reproduction. ### Captum BLEU: 68.8. Result ID: 698c5a2137894ef7d9cdbe1b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"Llama-2 (7B-Chat)"}. This record does not assert independent reproduction. ### Attention BLEU: 68.5. Result ID: 698b852e7a967eddb76f42ea. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"Vicuna (7B)"}. This record does not assert independent reproduction. ### Last-Attention BLEU: 68.4. Result ID: 698b852e7a967eddb76f42e9. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"Vicuna (7B)"}. This record does not assert independent reproduction. ### Integrated-Gradient BLEU: 67.2. Result ID: 698b852e7a967eddb76f42ec. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"Vicuna (7B)"}. This record does not assert independent reproduction. ### Attention BLEU: 66.5. Result ID: 698b852e7a967eddb76f42e4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"LlaMA-2 (7B-Chat)"}. This record does not assert independent reproduction. ### Attention BLEU: 66.4. Result ID: 698c5a2137894ef7d9cdbe1a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"Llama-2 (7B-Chat)"}. This record does not assert independent reproduction. ### Last-Attention BLEU: 66.2. Result ID: 698c5a2137894ef7d9cdbe19. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"Llama-2 (7B-Chat)"}. This record does not assert independent reproduction. ### Last-Attention BLEU: 66. Result ID: 698b852e7a967eddb76f42e3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"LlaMA-2 (7B-Chat)"}. This record does not assert independent reproduction. ### Integrated-Gradient BLEU: 64.6. Result ID: 698b852e7a967eddb76f42e6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"LlaMA-2 (7B-Chat)"}. This record does not assert independent reproduction. ### Captum BLEU: 64. Result ID: 698b852e7a967eddb76f42e5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"LlaMA-2 (7B-Chat)"}. This record does not assert independent reproduction. ### Integrated-Gradient BLEU: 61. Result ID: 698c5a2137894ef7d9cdbe1c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"Llama-2 (7B-Chat)"}. This record does not assert independent reproduction. ### Captum BLEU: 57.9. Result ID: 698b852e7a967eddb76f42eb. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"Vicuna (7B)"}. This record does not assert independent reproduction. ### JoPA BLEU: 57.9. Result ID: 698c5a2137894ef7d9cdbe1d. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"Llama-2 (7B-Chat)"}. This record does not assert independent reproduction. ### JoPA BLEU: 57.5. Result ID: 698b852e7a967eddb76f42e7. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"LlaMA-2 (7B-Chat)"}. This record does not assert independent reproduction. ### JoPA BLEU: 57.5. Result ID: 698b852e7a967eddb76f42ed. [Reporting paper]() [Method origin paper]() [Implementation]() 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.