# Understanding Finetuning for Factual Knowledge Extraction

arXiv: https://arxiv.org/abs/2406.14785 · HTML: https://mlsanity.com/paper/2406.14785/

## Key facts

- Cited by 32 papers in the map (data through 14 Oct 2026).
- Builds on 16 of its 28 references that are in the map.
- First posted on arXiv 20 Jun 2024.

## Builds on (most cited first, 16 of 16)

- [Measuring Massive Multitask Language Understanding](https://mlsanity.com/paper/2009.03300/) (2020)
- [Language Models as Knowledge Bases?](https://mlsanity.com/paper/1909.01066/) (2019)
- [Inference-Time Intervention: Eliciting Truthful Answers from a Language Model](https://mlsanity.com/paper/2306.03341/) (2023)
- [How Can We Know What Language Models Know?](https://mlsanity.com/paper/1911.12543/) (2019)
- [Discovering Latent Knowledge in Language Models Without Supervision](https://mlsanity.com/paper/2212.03827/) (2022)
- [How Much Knowledge Can You Pack Into the Parameters of a Language Model?](https://mlsanity.com/paper/2002.08910/) (2020)
- [Dissecting Recall of Factual Associations in Auto-Regressive Language Models](https://mlsanity.com/paper/2304.14767/) (2023)
- [Physics of Language Models: Part 3.1, Knowledge Storage and Extraction](https://mlsanity.com/paper/2309.14316/) (2023)
- [Impact of Pretraining Term Frequencies on Few-Shot Reasoning](https://mlsanity.com/paper/2202.07206/) (2022)
- [Sudden Drops in the Loss: Syntax Acquisition, Phase Transitions, and Simplicity Bias in MLMs](https://mlsanity.com/paper/2309.07311/) (2023)
- [Scan and Snap: Understanding Training Dynamics and Token Composition in 1-layer Transformer](https://mlsanity.com/paper/2305.16380/) (2023)
- [Unfamiliar Finetuning Examples Control How Language Models Hallucinate](https://mlsanity.com/paper/2403.05612/) (2024)
- [Exposing Attention Glitches with Flip-Flop Language Modeling](https://mlsanity.com/paper/2306.00946/) (2023)
- [Attention Satisfies: A Constraint-Satisfaction Lens on Factual Errors of Language Models](https://mlsanity.com/paper/2309.15098/) (2023)
- [Cognitive Dissonance: Why Do Language Model Outputs Disagree with Internal Representations of Truthfulness?](https://mlsanity.com/paper/2312.03729/) (2023)
- [Understanding Finetuning for Factual Knowledge Extraction from Language Models](https://mlsanity.com/paper/2301.11293/) (2023)

## Built on by (most cited first, 30 of 32)

- [A Survey of Self-Evolving Agents: What, When, How, and Where to Evolve on the Path to Artificial Super Intelligence](https://mlsanity.com/paper/2507.21046/) (2025)
- [Understanding Factual Recall in Transformers via Associative Memories](https://mlsanity.com/paper/2412.06538/) (2024)
- [Principled Data Selection for Alignment: The Hidden Risks of Difficult Examples](https://mlsanity.com/paper/2502.09650/) (2025)
- [Taming Knowledge Conflicts in Language Models](https://mlsanity.com/paper/2503.10996/) (2025)
- [Do Large Language Models Perform Latent Multi-Hop Reasoning without Exploiting Shortcuts?](https://mlsanity.com/paper/2411.16679/) (2024)
- [Upweighting Easy Samples in Fine-Tuning Mitigates Forgetting](https://mlsanity.com/paper/2502.02797/) (2025)
- [Data Mixing Can Induce Phase Transitions in Knowledge Acquisition](https://mlsanity.com/paper/2505.18091/) (2025)
- [Improving Model Factuality with Fine-grained Critique-based Evaluator](https://mlsanity.com/paper/2410.18359/) (2024)
- [Learning Facts at Scale with Active Reading](https://mlsanity.com/paper/2508.09494/) (2025)
- [Replaying pre-training data improves fine-tuning](https://mlsanity.com/paper/2603.04964/) (2026)
- [Context-Parametric Inversion: Why Instruction Finetuning Can Worsen Context Reliance](https://mlsanity.com/paper/2410.10796/) (2024)
- [ChroKnowledge: Unveiling Chronological Knowledge of Language Models in Multiple Domains](https://mlsanity.com/paper/2410.09870/) (2024)
- [From Style to Facts: Mapping the Boundaries of Knowledge Injection with Finetuning](https://mlsanity.com/paper/2503.05919/) (2025)
- [Evaluating Evaluation Metrics -- The Mirage of Hallucination Detection](https://mlsanity.com/paper/2504.18114/) (2025)
- [SHARP: Unlocking Interactive Hallucination via Stance Transfer in Role-Playing LLMs](https://mlsanity.com/paper/2411.07965/) (2024)
- [Rote Learning Considered Useful: Generalizing over Memorized Data in LLMs](https://mlsanity.com/paper/2507.21914/) (2025)
- [The Curious Case of Factuality Finetuning: Models' Internal Beliefs Can Improve Factuality](https://mlsanity.com/paper/2507.08371/) (2025)
- [Understanding Fact Recall in Language Models: Why Two-Stage Training Encourages Memorization but Mixed Training Teaches Knowledge](https://mlsanity.com/paper/2505.16178/) (2025)
- [From Parameters to Prompts: Understanding and Mitigating the Factuality Gap between Fine-Tuned LLMs](https://mlsanity.com/paper/2505.23410/) (2025)
- [From Memorization to Generalization: Fine-Tuning Large Language Models for Biomedical Term-to-Identifier Normalization](https://mlsanity.com/paper/2510.19036/) (2025)
- [REFLEX: Self-Refining Explainable Fact-Checking via Verdict-Anchored Style Control](https://mlsanity.com/paper/2511.20233/) (2025)
- [The Future of Facts: Tracing the Factual Generation-Verification Gap](https://mlsanity.com/paper/2605.27564/) (2026)
- [Fine-Tuning Dynamics of In-Context Factual Recall in Transformers](https://mlsanity.com/paper/2605.27774/) (2026)
- [Position: Pause Recycling LoRAs and Prioritize Mechanisms to Uncover Limits and Effectiveness](https://mlsanity.com/paper/2506.13479/) (2025)
- [Understanding New-Knowledge-Induced Factual Hallucinations in LLMs: Analysis and Interpretation](https://mlsanity.com/paper/2511.02626/) (2025)
- [Latent Knowledge as a Predictor of Fact Acquisition in Fine-Tuned Large Language Models](https://mlsanity.com/paper/2601.18468/) (2026)
- [Feature Resemblance: Towards a Theoretical Understanding of Analogical Reasoning in Transformers](https://mlsanity.com/paper/2603.05143/) (2026)
- [PRISM: Probability Reallocation with In-Span Masking for Knowledge-Sensitive Alignment](https://mlsanity.com/paper/2604.01682/) (2026)
- [What Models Know, How Well They Know It: Knowledge-Weighted Fine-Tuning for Learning When to Say "I Don't Know"](https://mlsanity.com/paper/2604.05779/) (2026)
- [Does Finetuning with Scientific Data Increase Hallucinations? A Multi-domain Factuality Evaluation of LLMs](https://mlsanity.com/paper/2606.21359/) (2026)

## Abstract (arXiv)

In this work, we study the impact of QA fine-tuning data on downstream factuality. We show that fine-tuning on lesser-known facts that are poorly stored during pretraining yields significantly worse factuality than fine-tuning on well-known facts, even when all facts are seen during pretraining. We prove this phenomenon theoretically, showing that training on lesser-known facts can lead the model to ignore subject entity names and instead output a generic plausible response even when the relevant factual knowledge is encoded in the model. On three question answering benchmarks (PopQA, Entity Questions, and MMLU) and two language models (Llama-2-7B and Mistral-7B), we find that (i) finetuning on a completely factual but lesser-known subset of the data deteriorates downstream factuality (5-10%) and (ii) finetuning on a subset of better-known examples matches or outperforms finetuning on the entire dataset. Ultimately, our results shed light on the interaction between pretrained knowledge and finetuning data and demonstrate the importance of taking into account how facts are stored in the pretrained model when fine-tuning for knowledge-intensive tasks.