From Continuous Pretraining to Domain-Adaptive Reranking via Task Vector Adaptation
Open Access
- 19 July 2026
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 3652-3657
Abstract
No abstract availableFunding Information
- Institute of Information & communications Technology Planning & Evaluation (IITP) (RS-2023-00216011)
- Institute of Information & communications Technology Planning & Evaluation (IITP) (RS-2024-00338140)
This publication has 10 references indexed in Scilit:
- Investigating Task Arithmetic for Zero-Shot Information Retrieval2025
- Layer-Aware Task Arithmetic: Disentangling Task-Specific and Instruction-Following Knowledge2025
- Fine-Tuning LLaMA for Multi-Stage Text Retrieval2024
- MultiLegalPile: A 689GB Multilingual Legal Corpus2024
- Chat Vector: A Simple Approach to Equip LLMs with Instruction Following and Model Alignment in New Languages2024
- DUQGen: Effective Unsupervised Domain Adaptation of Neural Rankers by Diversifying Synthetic Query Generation2024
- TIES-Merging: Resolving Interference When Merging Models2023
- Evaluating Unsupervised Text Classification: Zero-shot and Similarity-based Approaches2022
- Natural Questions: A Benchmark for Question Answering ResearchTransactions of the Association for Computational Linguistics, 2019
- Semantic Parsing on Freebase from Question-Answer Pairs2013