Causal-Guided Active Learning for Debiasing Large Language Models
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_ | 1866912008193966080 |
|---|---|
| author | Du, Li Sun, Zhouhao Ding, Xiao Ma, Yixuan Zhao, Yang Qiu, Kaitao Liu, Ting Qin, Bing |
| author_facet | Du, Li Sun, Zhouhao Ding, Xiao Ma, Yixuan Zhao, Yang Qiu, Kaitao Liu, Ting Qin, Bing |
| contents | Although achieving promising performance, recent analyses show that current generative large language models (LLMs) may still capture dataset biases and utilize them for generation, leading to poor generalizability and harmfulness of LLMs. However, due to the diversity of dataset biases and the over-optimization problem, previous prior-knowledge-based debiasing methods and fine-tuning-based debiasing methods may not be suitable for current LLMs. To address this issue, we explore combining active learning with the causal mechanisms and propose a casual-guided active learning (CAL) framework, which utilizes LLMs itself to automatically and autonomously identify informative biased samples and induce the bias patterns. Then a cost-effective and efficient in-context learning based method is employed to prevent LLMs from utilizing dataset biases during generation. Experimental results show that CAL can effectively recognize typical biased instances and induce various bias patterns for debiasing LLMs. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2408_12942 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Causal-Guided Active Learning for Debiasing Large Language Models Du, Li Sun, Zhouhao Ding, Xiao Ma, Yixuan Zhao, Yang Qiu, Kaitao Liu, Ting Qin, Bing Computation and Language Artificial Intelligence Although achieving promising performance, recent analyses show that current generative large language models (LLMs) may still capture dataset biases and utilize them for generation, leading to poor generalizability and harmfulness of LLMs. However, due to the diversity of dataset biases and the over-optimization problem, previous prior-knowledge-based debiasing methods and fine-tuning-based debiasing methods may not be suitable for current LLMs. To address this issue, we explore combining active learning with the causal mechanisms and propose a casual-guided active learning (CAL) framework, which utilizes LLMs itself to automatically and autonomously identify informative biased samples and induce the bias patterns. Then a cost-effective and efficient in-context learning based method is employed to prevent LLMs from utilizing dataset biases during generation. Experimental results show that CAL can effectively recognize typical biased instances and induce various bias patterns for debiasing LLMs. |
| title | Causal-Guided Active Learning for Debiasing Large Language Models |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2408.12942 |