IKnow: Instruction-Knowledge-Aware Continual Pretraining for Effective Domain Adaptation

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Zhang, Tianyi, Mai, Florian, Flek, Lucie
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866914109100916736
author Zhang, Tianyi
Mai, Florian
Flek, Lucie
author_facet Zhang, Tianyi
Mai, Florian
Flek, Lucie
contents Continual pretraining promises to adapt large language models (LLMs) to new domains using only unlabeled test-time data, but naively applying standard self-supervised objectives to instruction-tuned models is known to degrade their instruction-following capability and semantic representations. Existing fixes assume access to the original base model or rely on knowledge from an external domain-specific database - both of which pose a realistic barrier in settings where the base model weights are withheld for safety reasons or reliable external corpora are unavailable. In this work, we propose Instruction-Knowledge-Aware Continual Adaptation (IKnow), a simple and general framework that formulates novel self-supervised objectives in the instruction-response dialogue format. Rather than depend- ing on external resources, IKnow leverages domain knowledge embedded within the text itself and learns to encode it at a deeper semantic level.
format Preprint
id arxiv_https___arxiv_org_abs_2510_20377
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle IKnow: Instruction-Knowledge-Aware Continual Pretraining for Effective Domain Adaptation
Zhang, Tianyi
Mai, Florian
Flek, Lucie
Artificial Intelligence
Computation and Language
Continual pretraining promises to adapt large language models (LLMs) to new domains using only unlabeled test-time data, but naively applying standard self-supervised objectives to instruction-tuned models is known to degrade their instruction-following capability and semantic representations. Existing fixes assume access to the original base model or rely on knowledge from an external domain-specific database - both of which pose a realistic barrier in settings where the base model weights are withheld for safety reasons or reliable external corpora are unavailable. In this work, we propose Instruction-Knowledge-Aware Continual Adaptation (IKnow), a simple and general framework that formulates novel self-supervised objectives in the instruction-response dialogue format. Rather than depend- ing on external resources, IKnow leverages domain knowledge embedded within the text itself and learns to encode it at a deeper semantic level.
title IKnow: Instruction-Knowledge-Aware Continual Pretraining for Effective Domain Adaptation
topic Artificial Intelligence
Computation and Language
url https://arxiv.org/abs/2510.20377