# Inducing Epistemological Humility in Large Language Models: A Targeted SFT Approach to Reducing Hallucination

arXiv: https://arxiv.org/abs/2603.17504 · HTML: https://mlsanity.com/paper/2603.17504/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 24 of its 31 references that are in the map.
- First posted on arXiv 18 Mar 2026.

## Builds on (most cited first, 24 of 24)

- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [Measuring Massive Multitask Language Understanding](https://mlsanity.com/paper/2009.03300/) (2020)
- [Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback](https://mlsanity.com/paper/2204.05862/) (2022)
- [Understanding intermediate layers using linear classifier probes](https://mlsanity.com/paper/1610.01644/) (2016)
- [Representation Engineering: A Top-Down Approach to AI Transparency](https://mlsanity.com/paper/2310.01405/) (2023)
- [FActScore: Fine-grained Atomic Evaluation of Factual Precision in Long Form Text Generation](https://mlsanity.com/paper/2305.14251/) (2023)
- [Enhancing Chat Language Models by Scaling High-quality Instructional Conversations](https://mlsanity.com/paper/2305.14233/) (2023)
- [The Geometry of Truth: Emergent Linear Structure in Large Language Model Representations of True/False Datasets](https://mlsanity.com/paper/2310.06824/) (2023)
- [Dolma: an Open Corpus of Three Trillion Tokens for Language Model Pretraining Research](https://mlsanity.com/paper/2402.00159/) (2024)
- [Reasoning Models Don't Always Say What They Think](https://mlsanity.com/paper/2505.05410/) (2025)
- [What Makes Good Data for Alignment? A Comprehensive Study of Automatic Data Selection in Instruction Tuning](https://mlsanity.com/paper/2312.15685/) (2023)
- [Know Your Limits: A Survey of Abstention in Large Language Models](https://mlsanity.com/paper/2407.18418/) (2024)
- [LoRA vs Full Fine-tuning: An Illusion of Equivalence](https://mlsanity.com/paper/2410.21228/) (2024)
- [The CoT Collection: Improving Zero-shot and Few-shot Learning of Language Models via Chain-of-Thought Fine-Tuning](https://mlsanity.com/paper/2305.14045/) (2023)
- [CoEdIT: Text Editing by Task-Specific Instruction Tuning](https://mlsanity.com/paper/2305.09857/) (2023)
- [AILuminate: Introducing v1.0 of the AI Risk and Reliability Benchmark from MLCommons](https://mlsanity.com/paper/2503.05731/) (2025)
- [MUFFIN: Curating Multi-Faceted Instructions for Improving Instruction-Following](https://mlsanity.com/paper/2312.02436/) (2023)
- [Know When To Stop: A Study of Semantic Drift in Text Generation](https://mlsanity.com/paper/2404.05411/) (2024)
- [Know the Unknown: An Uncertainty-Sensitive Method for LLM Instruction Tuning](https://mlsanity.com/paper/2406.10099/) (2024)
- [Fine-Tuning Large Language Models to Appropriately Abstain with Semantic Entropy](https://mlsanity.com/paper/2410.17234/) (2024)
- [HypoTermQA: Hypothetical Terms Dataset for Benchmarking Hallucination Tendency of LLMs](https://mlsanity.com/paper/2402.16211/) (2024)
- [Balancing Truthfulness and Informativeness with Uncertainty-Aware Instruction Fine-Tuning](https://mlsanity.com/paper/2502.11962/) (2025)

## Abstract (arXiv)

Large language models (LLMs) often hallucinate, producing fluent but false information, partly because supervised fine-tuning (SFT) implicitly rewards always responding. We introduce textitHypoTermInstruct, an SFT dataset (31,487 responses for 11,151 questions) designed to teach models epistemological humility-the ability to recognize the limits of their own knowledge and admit uncertainty. This is achieved through questions about non-existent "hypothetical" terms. We also release textitHypoTermQA-Enhanced, a benchmark for hallucination tendency strengthened through multiple validations. We conducted 800 controlled LoRA SFT runs across textitLlama3.1-8B and textitGemma3-4B (base and instruct), testing 100 fine-tuning configurations with paired controls. Our results demonstrate that replacing generic instruction data with textitHypoTermInstruct significantly improves the HypoTerm Score (median increases of 0.19% to 25.91%) and FactScore (+0.39% to +0.86%), while maintaining stable performance on MMLU (minimal decreases of 0.26% to 0.35%). Our work demonstrates that targeted, high-quality SFT data teaching meta-cognitive skills can effectively reduce hallucination without preference/RL pipelines, providing mechanistic insights and a practical path toward more reliable AI systems.