TY - GEN
T1 - TELLME
T2 - 19th Conference of the European Chapter of the Association for Computational Linguistics, Findings of EACL 2026
AU - Kim, Minjun
AU - Won, Inho
AU - Lim, Hyeonseok
AU - Kim, Min Kyu
AU - Yuk, Junghun
AU - Go, Wooyoung
AU - Park, Jongyoul
AU - Park, Jungyeul
AU - Lim, Kyung Tae
N1 - Publisher Copyright:
©2026 Association for Computational Linguistics.
PY - 2026
Y1 - 2026
N2 - Continual pre-training (CPT) has been widely adopted as a method for domain adaptation in large language models. However, CPT has consistently been accompanied by challenges, such as the difficulty of acquiring large-scale domain-specific datasets and high computational costs. In this study, we propose a novel method called Test-Enhanced Learning for Language Model Enrichment (TELLME) to alleviate these issues. TELLME leverages the Test-Enhanced Learning (TEL) principle, whereby the model’s training efficiency is improved using quizzes during training. It integrates this principle with CPT, thereby promoting efficient domain-specific knowledge acquisition and long-term memory retention. Experimental results demonstrate that TELLME outperforms existing methods by up to 23.6% in the financial domain and achieves a 9.8% improvement in long-term memory retention. The model and TELLME dataset are available at huggingface.co/anonymous4459.
AB - Continual pre-training (CPT) has been widely adopted as a method for domain adaptation in large language models. However, CPT has consistently been accompanied by challenges, such as the difficulty of acquiring large-scale domain-specific datasets and high computational costs. In this study, we propose a novel method called Test-Enhanced Learning for Language Model Enrichment (TELLME) to alleviate these issues. TELLME leverages the Test-Enhanced Learning (TEL) principle, whereby the model’s training efficiency is improved using quizzes during training. It integrates this principle with CPT, thereby promoting efficient domain-specific knowledge acquisition and long-term memory retention. Experimental results demonstrate that TELLME outperforms existing methods by up to 23.6% in the financial domain and achieves a 9.8% improvement in long-term memory retention. The model and TELLME dataset are available at huggingface.co/anonymous4459.
UR - https://www.scopus.com/pages/publications/105038918968
U2 - 10.18653/v1/2026.findings-eacl.84
DO - 10.18653/v1/2026.findings-eacl.84
M3 - Conference contribution
AN - SCOPUS:105038918968
T3 - 19th Conference of the European Chapter of the Association for Computational Linguistics, Findings of EACL 2026
SP - 1655
EP - 1677
BT - 19th Conference of the European Chapter of the Association for Computational Linguistics, Findings of EACL 2026
PB - Association for Computational Linguistics (ACL)
Y2 - 24 March 2026 through 29 March 2026
ER -