Self-Alignment for Factuality: Mitigating Hallucinations in LLMs via Self-Evaluation
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_ | 1866929381511790592 |
|---|---|
| author | Zhang, Xiaoying Peng, Baolin Tian, Ye Zhou, Jingyan Jin, Lifeng Song, Linfeng Mi, Haitao Meng, Helen |
| author_facet | Zhang, Xiaoying Peng, Baolin Tian, Ye Zhou, Jingyan Jin, Lifeng Song, Linfeng Mi, Haitao Meng, Helen |
| contents | Despite showing increasingly human-like abilities, large language models (LLMs) often struggle with factual inaccuracies, i.e. "hallucinations", even when they hold relevant knowledge. To address these hallucinations, current approaches typically necessitate high-quality human factuality annotations. In this work, we explore Self-Alignment for Factuality, where we leverage the self-evaluation capability of an LLM to provide training signals that steer the model towards factuality. Specifically, we incorporate Self-Eval, a self-evaluation component, to prompt an LLM to validate the factuality of its own generated responses solely based on its internal knowledge. Additionally, we design Self-Knowledge Tuning (SK-Tuning) to augment the LLM's self-evaluation ability by improving the model's confidence estimation and calibration. We then utilize these self-annotated responses to fine-tune the model via Direct Preference Optimization algorithm. We show that the proposed self-alignment approach substantially enhances factual accuracy over Llama family models across three key knowledge-intensive tasks on TruthfulQA and BioGEN. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2402_09267 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Self-Alignment for Factuality: Mitigating Hallucinations in LLMs via Self-Evaluation Zhang, Xiaoying Peng, Baolin Tian, Ye Zhou, Jingyan Jin, Lifeng Song, Linfeng Mi, Haitao Meng, Helen Computation and Language Artificial Intelligence Despite showing increasingly human-like abilities, large language models (LLMs) often struggle with factual inaccuracies, i.e. "hallucinations", even when they hold relevant knowledge. To address these hallucinations, current approaches typically necessitate high-quality human factuality annotations. In this work, we explore Self-Alignment for Factuality, where we leverage the self-evaluation capability of an LLM to provide training signals that steer the model towards factuality. Specifically, we incorporate Self-Eval, a self-evaluation component, to prompt an LLM to validate the factuality of its own generated responses solely based on its internal knowledge. Additionally, we design Self-Knowledge Tuning (SK-Tuning) to augment the LLM's self-evaluation ability by improving the model's confidence estimation and calibration. We then utilize these self-annotated responses to fine-tune the model via Direct Preference Optimization algorithm. We show that the proposed self-alignment approach substantially enhances factual accuracy over Llama family models across three key knowledge-intensive tasks on TruthfulQA and BioGEN. |
| title | Self-Alignment for Factuality: Mitigating Hallucinations in LLMs via Self-Evaluation |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2402.09267 |