@inproceedings{ruan-etal-2026-gift,
    title = "{GIFT}: Guided Fine-Tuning and Transfer for Enhancing Instruction-Tuned Language Models",
    author = "Ruan, Zhiwen  and
      Du, Yichao  and
      Zheng, Jianjie  and
      Wang, Longyue  and
      Chen, Yun  and
      Li, Peng  and
      Su, Jinsong  and
      Liu, Yang  and
      Chen, Guanhua",
    editor = "Liakata, Maria  and
      Moreira, Viviane P.  and
      Zhang, Jiajun  and
      Jurgens, David",
    booktitle = "Proceedings of the 64th Annual Meeting of the {A}ssociation for {C}omputational {L}inguistics (Volume 1: Long Papers)",
    month = jul,
    year = "2026",
    address = "San Diego, California, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-acl/2026.acl-long.358/",
    pages = "7866--7878",
    ISBN = "979-8-89176-390-6",
    abstract = "Instruction-tuned large language models (LLMs) exhibit strong instruction-following and generalization abilities, enabled by expensive post-training pipelines. However, adapting them to specific downstream tasks remains challenging: direct fine-tuning often disrupts this delicate balance, while existing adapter-based transfer methods typically treat the instruction-tuned model as a passive target that only participates at the final merging stage. We propose GIFT (Guided Fine-Tuning and Transfer), a simple and efficient framework that incorporates instruction-level guidance into task adaptation. GIFT fine-tunes a low-rank adapter on the pretrained base model using token-level confidence signals derived from the instruction-tuned model. The learned adapter is then merged into the instruction-tuned model, yielding task-specialized models that preserve general instruction-following behavior. We evaluate GIFT on mathematical reasoning and knowledge-intensive benchmarks across multiple model families and scales. Results show that GIFT consistently outperforms direct fine-tuning and representative transfer-based baselines, while maintaining robust generalization and favorable test-time scaling behavior."
}