Linguistic Rule Induction Improves Adversarial and OOD Robustness in Large Language Models

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216208%3A11320%2F25%3ATDED4A2K" target="_blank" >RIV/00216208:11320/25:TDED4A2K - isvavai.cz</a>

Výsledek na webu

<a href="https://www.scopus.com/inward/record.uri?eid=2-s2.0-85195915375&partnerID=40&md5=16d68f89b42f03ed02b1e1775dbf6a57" target="_blank" >https://www.scopus.com/inward/record.uri?eid=2-s2.0-85195915375&partnerID=40&md5=16d68f89b42f03ed02b1e1775dbf6a57</a>

DOI - Digital Object Identifier

—

Jazyk výsledku

angličtina

Název v původním jazyce

Linguistic Rule Induction Improves Adversarial and OOD Robustness in Large Language Models

Popis výsledku v původním jazyce

Ensuring robustness is especially important when AI is deployed in responsible or safety-critical environments. ChatGPT can perform brilliantly in both adversarial and out-of-distribution (OOD) robustness. Still, other popular large language models (LLMs), like LLaMA-2, ERNIE, and ChatGLM, do not perform satisfactorily in this regard. Therefore, it is valuable to study what efforts play essential roles in ChatGPT, and how to transfer these efforts to other LLMs. This paper experimentally finds that linguistic rule induction is the foundation for identifying the cause-effect relationships in LLMs. Accurately processing the cause-effect relationships in LLMs can improve their adversarial and OOD robustness. Furthermore, we explore a low-cost way of aligning LLMs with linguistic rules. Specifically, we constructed a linguistic rule instruction dataset to fine-tune LLMs. To further energize LLMs for reasoning step-by-step with the linguistic rules, we propose the task-relevant LingR-based chain-of-thoughts. Experiments showed that LingR-induced LLaMA-13B achieves comparable or better results with GPT-3.5 and GPT-4 on various adversarial and OOD robustness evaluations. © 2024 ELRA Language Resource Association: CC BY-NC 4.0.

Název v anglickém jazyce

Linguistic Rule Induction Improves Adversarial and OOD Robustness in Large Language Models

Popis výsledku anglicky

Ensuring robustness is especially important when AI is deployed in responsible or safety-critical environments. ChatGPT can perform brilliantly in both adversarial and out-of-distribution (OOD) robustness. Still, other popular large language models (LLMs), like LLaMA-2, ERNIE, and ChatGLM, do not perform satisfactorily in this regard. Therefore, it is valuable to study what efforts play essential roles in ChatGPT, and how to transfer these efforts to other LLMs. This paper experimentally finds that linguistic rule induction is the foundation for identifying the cause-effect relationships in LLMs. Accurately processing the cause-effect relationships in LLMs can improve their adversarial and OOD robustness. Furthermore, we explore a low-cost way of aligning LLMs with linguistic rules. Specifically, we constructed a linguistic rule instruction dataset to fine-tune LLMs. To further energize LLMs for reasoning step-by-step with the linguistic rules, we propose the task-relevant LingR-based chain-of-thoughts. Experiments showed that LingR-induced LLaMA-13B achieves comparable or better results with GPT-3.5 and GPT-4 on various adversarial and OOD robustness evaluations. © 2024 ELRA Language Resource Association: CC BY-NC 4.0.

Druh

D - Stať ve sborníku

CEP obor

—

OECD FORD obor

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Projekt

—

Návaznosti

—

Rok uplatnění

2024

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název statě ve sborníku

Jt. Int. Conf. Comput. Linguist., Lang. Resour. Eval., LREC-COLING - Main Conf. Proc.

ISBN

978-249381410-4

ISSN

—

e-ISSN

—

Počet stran výsledku

13

Strana od-do

10565-10577

Název nakladatele

European Language Resources Association (ELRA)

Místo vydání

—

Místo konání akce

Torino, Italia

Datum konání akce

1. 1. 2025

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

—