# Balancing Truthfulness and Informativeness with Uncertainty-Aware Instruction Fine-Tuning

arXiv: https://arxiv.org/abs/2502.11962 · HTML: https://mlsanity.com/paper/2502.11962/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Builds on 22 of its 31 references that are in the map.
- First posted on arXiv 17 Feb 2025.

## Builds on (most cited first, 22 of 22)

- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena](https://mlsanity.com/paper/2306.05685/) (2023)
- [Efficient Memory Management for Large Language Model Serving with PagedAttention](https://mlsanity.com/paper/2309.06180/) (2023)
- [Decoupled Weight Decay Regularization](https://mlsanity.com/paper/1711.05101/) (2017)
- [FActScore: Fine-grained Atomic Evaluation of Factual Precision in Long Form Text Generation](https://mlsanity.com/paper/2305.14251/) (2023)
- [LIMA: Less Is More for Alignment](https://mlsanity.com/paper/2305.11206/) (2023)
- [Can LLMs Express Their Uncertainty? An Empirical Evaluation of Confidence Elicitation in LLMs](https://mlsanity.com/paper/2306.13063/) (2023)
- [What Makes Good Data for Alignment? A Comprehensive Study of Automatic Data Selection in Instruction Tuning](https://mlsanity.com/paper/2312.15685/) (2023)
- [Magpie: Alignment Data Synthesis from Scratch by Prompting Aligned LLMs with Nothing](https://mlsanity.com/paper/2406.08464/) (2024)
- [Does Fine-Tuning LLMs on New Knowledge Encourage Hallucinations?](https://mlsanity.com/paper/2405.05904/) (2024)
- [Fact-Checking the Output of Large Language Models via Token-Level Uncertainty Quantification](https://mlsanity.com/paper/2403.04696/) (2024)
- [AlpaCare:Instruction-tuned Large Language Models for Medical Application](https://mlsanity.com/paper/2310.14558/) (2023)
- [Unfamiliar Finetuning Examples Control How Language Models Hallucinate](https://mlsanity.com/paper/2403.05612/) (2024)
- [Long Is More for Alignment: A Simple but Tough-to-Beat Baseline for Instruction Fine-Tuning](https://mlsanity.com/paper/2402.04833/) (2024)
- [Rejection Improves Reliability: Training LLMs to Refuse Unknown Questions Using RL from Knowledge Feedback](https://mlsanity.com/paper/2403.18349/) (2024)
- [Linguistic Calibration of Long-Form Generations](https://mlsanity.com/paper/2404.00474/) (2024)
- [RAG-QA Arena: Evaluating Domain Robustness for Long-form Retrieval Augmented Question Answering](https://mlsanity.com/paper/2407.13998/) (2024)
- [WildHallucinations: Evaluating Long-form Factuality in LLMs with Real-World Entity Queries](https://mlsanity.com/paper/2407.17468/) (2024)
- [LoGU: Long-form Generation with Uncertainty Expressions](https://mlsanity.com/paper/2410.14309/) (2024)
- [LawInstruct: A Resource for Studying Language Model Adaptation to the Legal Domain](https://mlsanity.com/paper/2404.02127/) (2024)
- [A Comparative Analysis of Instruction Fine-Tuning LLMs for Financial Text Classification](https://mlsanity.com/paper/2411.02476/) (2024)
- [SusGen-GPT: A Data-Centric LLM for Financial NLP and Sustainability Report Generation](https://mlsanity.com/paper/2412.10906/) (2024)

## Built on by (most cited first, 3 of 3)

- [Train for Truth, Keep the Skills: Binary Retrieval-Augmented Reward Mitigates Hallucinations](https://mlsanity.com/paper/2510.17733/) (2025)
- [Granuscore: A Reference-Free Measure of Granularity for Text Analysis and Question Answering](https://mlsanity.com/paper/2605.26620/) (2026)
- [Inducing Epistemological Humility in Large Language Models: A Targeted SFT Approach to Reducing Hallucination](https://mlsanity.com/paper/2603.17504/) (2026)

## Abstract (arXiv)

Instruction fine-tuning (IFT) can increase the informativeness of large language models (LLMs), but may reduce their truthfulness. This trade-off arises because IFT steers LLMs to generate responses containing long-tail knowledge that was not well covered during pre-training. As a result, models become more informative but less accurate when generalizing to unseen tasks. In this paper, we empirically demonstrate how unfamiliar knowledge in IFT datasets can negatively affect the truthfulness of LLMs, and we introduce two new IFT paradigms, UNITcut and UNITref, to address this issue. UNITcut identifies and removes unfamiliar knowledge from IFT datasets to mitigate its impact on model truthfulness, whereas UNITref trains LLMs to recognize their uncertainty and explicitly indicate it at the end of their responses. Our experiments show that UNITcut substantially improves LLM truthfulness, while UNITref maintains high informativeness and reduces hallucinations by distinguishing between confident and uncertain statements.