Measuring and Enhancing Trustworthiness of LLMs in RAG through Grounded Attributions and Learning to Refuse
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866908336706813952 |
|---|---|
| author | Song, Maojia Sim, Shang Hong Bhardwaj, Rishabh Chieu, Hai Leong Majumder, Navonil Poria, Soujanya |
| author_facet | Song, Maojia Sim, Shang Hong Bhardwaj, Rishabh Chieu, Hai Leong Majumder, Navonil Poria, Soujanya |
| contents | LLMs are an integral component of retrieval-augmented generation (RAG) systems. While many studies focus on evaluating the overall quality of end-to-end RAG systems, there is a gap in understanding the appropriateness of LLMs for the RAG task. To address this, we introduce Trust-Score, a holistic metric that evaluates the trustworthiness of LLMs within the RAG framework. Our results show that various prompting methods, such as in-context learning, fail to effectively adapt LLMs to the RAG task as measured by Trust-Score. Consequently, we propose Trust-Align, a method to align LLMs for improved Trust-Score performance. 26 out of 27 models aligned using Trust-Align substantially outperform competitive baselines on ASQA, QAMPARI, and ELI5. Specifically, in LLaMA-3-8b, Trust-Align outperforms FRONT on ASQA (up 12.56), QAMPARI (up 36.04), and ELI5 (up 17.69). Trust-Align also significantly enhances models' ability to correctly refuse and provide quality citations. We also demonstrate the effectiveness of Trust-Align across different open-weight models, including the LLaMA series (1b to 8b), Qwen-2.5 series (0.5b to 7b), and Phi3.5 (3.8b). We release our code at https://github.com/declare-lab/trust-align. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2409_11242 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Measuring and Enhancing Trustworthiness of LLMs in RAG through Grounded Attributions and Learning to Refuse Song, Maojia Sim, Shang Hong Bhardwaj, Rishabh Chieu, Hai Leong Majumder, Navonil Poria, Soujanya Computation and Language LLMs are an integral component of retrieval-augmented generation (RAG) systems. While many studies focus on evaluating the overall quality of end-to-end RAG systems, there is a gap in understanding the appropriateness of LLMs for the RAG task. To address this, we introduce Trust-Score, a holistic metric that evaluates the trustworthiness of LLMs within the RAG framework. Our results show that various prompting methods, such as in-context learning, fail to effectively adapt LLMs to the RAG task as measured by Trust-Score. Consequently, we propose Trust-Align, a method to align LLMs for improved Trust-Score performance. 26 out of 27 models aligned using Trust-Align substantially outperform competitive baselines on ASQA, QAMPARI, and ELI5. Specifically, in LLaMA-3-8b, Trust-Align outperforms FRONT on ASQA (up 12.56), QAMPARI (up 36.04), and ELI5 (up 17.69). Trust-Align also significantly enhances models' ability to correctly refuse and provide quality citations. We also demonstrate the effectiveness of Trust-Align across different open-weight models, including the LLaMA series (1b to 8b), Qwen-2.5 series (0.5b to 7b), and Phi3.5 (3.8b). We release our code at https://github.com/declare-lab/trust-align. |
| title | Measuring and Enhancing Trustworthiness of LLMs in RAG through Grounded Attributions and Learning to Refuse |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2409.11242 |