Instruction Fine-tuning (IFT) can enhance the helpfulness of Large Language Models (LLMs), but it may lower their truthfulness. This trade-off arises because IFT steers LLMs to generate responses with long-tail knowledge that is not well covered during pre-training, leading to more informative but less truthful answers when generalizing to unseen tasks. In this paper, we empirically demonstrate this helpfulness-truthfulness trade-off in IFT and propose , a novel IFT paradigm to address it. UNIT teaches LLMs to recognize their uncertainty and explicitly reflect it at the end of their responses. Experimental results show that UNIT-tuned models maintain their helpfulness while distinguishing between certain and uncertain claims, thereby reducing hallucinations.
View on arXiv@article{wu2025_2502.11962,
title={ Navigating the Helpfulness-Truthfulness Trade-Off with Uncertainty-Aware Instruction Fine-Tuning },
author={ Tianyi Wu and Jingwei Ni and Bryan Hooi and Jiaheng Zhang and Elliott Ash and See-Kiong Ng and Mrinmaya Sachan and Markus Leippold },
journal={arXiv preprint arXiv:2502.11962},
year={ 2025 }
}