FaithLM: Towards Faithful Explanations for Large Language Models
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2024
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866909870207270912 |
|---|---|
| author | Chuang, Yu-Neng Wang, Guanchu Chang, Chia-Yuan Tang, Ruixiang Zhong, Shaochen Yang, Fan Du, Mengnan Cai, Xuanting Braverman, Vladimir Hu, Xia |
| author_facet | Chuang, Yu-Neng Wang, Guanchu Chang, Chia-Yuan Tang, Ruixiang Zhong, Shaochen Yang, Fan Du, Mengnan Cai, Xuanting Braverman, Vladimir Hu, Xia |
| contents | Large language models (LLMs) increasingly produce natural language explanations, yet these explanations often lack faithfulness, and they do not reliably reflect the evidence the model uses to decide. We introduce FaithLM, a model-agnostic framework that evaluates and improves the faithfulness of LLM explanations without token masking or task-specific heuristics. FaithLM formalizes explanation faithfulness as an intervention property: a faithful explanation should yield a prediction shift when its content is contradicted. Theoretical analysis shows that the resulting contrary-hint score is a sound and discriminative estimator of faithfulness. Building on this principle, FaithLM iteratively refines both the elicitation prompt and the explanation to maximize the measured score. Experiments on three multi-domain datasets and multiple LLM backbones demonstrate that FaithLM consistently increases faithfulness and produces explanations more aligned with human rationales than strong self-explanation baselines. These findings highlight that intervention-based evaluation, coupled with iterative optimization, provides a principled route toward faithful and reliable LLM explanations. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2402_04678 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | FaithLM: Towards Faithful Explanations for Large Language Models Chuang, Yu-Neng Wang, Guanchu Chang, Chia-Yuan Tang, Ruixiang Zhong, Shaochen Yang, Fan Du, Mengnan Cai, Xuanting Braverman, Vladimir Hu, Xia Computation and Language Artificial Intelligence Machine Learning Large language models (LLMs) increasingly produce natural language explanations, yet these explanations often lack faithfulness, and they do not reliably reflect the evidence the model uses to decide. We introduce FaithLM, a model-agnostic framework that evaluates and improves the faithfulness of LLM explanations without token masking or task-specific heuristics. FaithLM formalizes explanation faithfulness as an intervention property: a faithful explanation should yield a prediction shift when its content is contradicted. Theoretical analysis shows that the resulting contrary-hint score is a sound and discriminative estimator of faithfulness. Building on this principle, FaithLM iteratively refines both the elicitation prompt and the explanation to maximize the measured score. Experiments on three multi-domain datasets and multiple LLM backbones demonstrate that FaithLM consistently increases faithfulness and produces explanations more aligned with human rationales than strong self-explanation baselines. These findings highlight that intervention-based evaluation, coupled with iterative optimization, provides a principled route toward faithful and reliable LLM explanations. |
| title | FaithLM: Towards Faithful Explanations for Large Language Models |
| topic | Computation and Language Artificial Intelligence Machine Learning |
| url | https://arxiv.org/abs/2402.04678 |