Salvato in:
Dettagli Bibliografici
Autori principali: Chu, Zhixuan, Wang, Yan, Li, Longfei, Wang, Zhibo, Qin, Zhan, Ren, Kui
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:https://arxiv.org/abs/2405.04160
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866910588893921280
author Chu, Zhixuan
Wang, Yan
Li, Longfei
Wang, Zhibo
Qin, Zhan
Ren, Kui
author_facet Chu, Zhixuan
Wang, Yan
Li, Longfei
Wang, Zhibo
Qin, Zhan
Ren, Kui
contents Large Language Models (LLMs) have shown impressive performance in natural language tasks, but their outputs can exhibit undesirable attributes or biases. Existing methods for steering LLMs toward desired attributes often assume unbiased representations and rely solely on steering prompts. However, the representations learned from pre-training can introduce semantic biases that influence the steering process, leading to suboptimal results. We propose LLMGuardrail, a novel framework that incorporates causal analysis and adversarial learning to obtain unbiased steering representations in LLMs. LLMGuardrail systematically identifies and blocks the confounding effects of biases, enabling the extraction of unbiased steering representations. Additionally, it includes an explainable component that provides insights into the alignment between the generated output and the desired direction. Experiments demonstrate LLMGuardrail's effectiveness in steering LLMs toward desired attributes while mitigating biases. Our work contributes to the development of safe and reliable LLMs that align with desired attributes.
format Preprint
id arxiv_https___arxiv_org_abs_2405_04160
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle A Causal Explainable Guardrails for Large Language Models
Chu, Zhixuan
Wang, Yan
Li, Longfei
Wang, Zhibo
Qin, Zhan
Ren, Kui
Computation and Language
Large Language Models (LLMs) have shown impressive performance in natural language tasks, but their outputs can exhibit undesirable attributes or biases. Existing methods for steering LLMs toward desired attributes often assume unbiased representations and rely solely on steering prompts. However, the representations learned from pre-training can introduce semantic biases that influence the steering process, leading to suboptimal results. We propose LLMGuardrail, a novel framework that incorporates causal analysis and adversarial learning to obtain unbiased steering representations in LLMs. LLMGuardrail systematically identifies and blocks the confounding effects of biases, enabling the extraction of unbiased steering representations. Additionally, it includes an explainable component that provides insights into the alignment between the generated output and the desired direction. Experiments demonstrate LLMGuardrail's effectiveness in steering LLMs toward desired attributes while mitigating biases. Our work contributes to the development of safe and reliable LLMs that align with desired attributes.
title A Causal Explainable Guardrails for Large Language Models
topic Computation and Language
url https://arxiv.org/abs/2405.04160