TELLME: Test-Enhanced Learning for Language Model Enrichment
TELLME:面向语言模型增强的测试增强学习
机构 * Korea Advanced Institute of Science and Technology(韩国科学技术院) ; Seoul National University of Science and Technology(首尔科学技术大学) ; National Security Research Institute(国家安全研究院)
AI总结 本研究提出TELLME方法,将测试增强学习(TEL)原理与持续预训练(CPT)结合,缓解大语言模型领域自适应中数据与成本问题,在金融领域表现优于现有方法,长期记忆保留提升显著。
Comments Findings of the Association for Computational Linguistics: EACL 2026
Journal ref Findings of the Association for Computational Linguistics: EACL 2026, pages 1655-1677