# FactBench: A Dynamic Benchmark for In-the-Wild Language Model Factuality Evaluation

arXiv: https://arxiv.org/abs/2410.22257 · HTML: https://mlsanity.com/paper/2410.22257/

## Key facts

- Cited by 25 papers in the map (data through 14 Oct 2026).
- Builds on 23 of its 77 references that are in the map.
- First posted on arXiv 29 Oct 2024.

## Builds on (most cited first, 23 of 23)

- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [Chain-of-Thought Prompting Elicits Reasoning in Large Language Models](https://mlsanity.com/paper/2201.11903/) (2022)
- [TruthfulQA: Measuring How Models Mimic Human Falsehoods](https://mlsanity.com/paper/2109.07958/) (2021)
- [UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction](https://mlsanity.com/paper/1802.03426/) (2018)
- [A Broad-Coverage Challenge Corpus for Sentence Understanding through Inference](https://mlsanity.com/paper/1704.05426/) (2017)
- [FActScore: Fine-grained Atomic Evaluation of Factual Precision in Long Form Text Generation](https://mlsanity.com/paper/2305.14251/) (2023)
- [DeBERTaV3: Improving DeBERTa using ELECTRA-Style Pre-Training with Gradient-Disentangled Embedding Sharing](https://mlsanity.com/paper/2111.09543/) (2021)
- [When Not to Trust Language Models: Investigating Effectiveness of Parametric and Non-Parametric Memories](https://mlsanity.com/paper/2212.10511/) (2022)
- [BERTopic: Neural topic modeling with a class-based TF-IDF procedure](https://mlsanity.com/paper/2203.05794/) (2022)
- [WildChat: 1M ChatGPT Interaction Logs in the Wild](https://mlsanity.com/paper/2405.01470/) (2024)
- [LMSYS-Chat-1M: A Large-Scale Real-World LLM Conversation Dataset](https://mlsanity.com/paper/2309.11998/) (2023)
- [HaluEval: A Large-Scale Hallucination Evaluation Benchmark for Large Language Models](https://mlsanity.com/paper/2305.11747/) (2023)
- [Data Contamination: From Memorization to Exploitation](https://mlsanity.com/paper/2203.08242/) (2022)
- [Long-form factuality in large language models](https://mlsanity.com/paper/2403.18802/) (2024)
- [VERISCORE: Evaluating the factuality of verifiable claims in long-form text generation](https://mlsanity.com/paper/2406.19276/) (2024)
- [Generating Benchmarks for Factuality Evaluation of Language Models](https://mlsanity.com/paper/2307.06908/) (2023)
- [ExpertQA: Expert-Curated Questions and Attributed Answers](https://mlsanity.com/paper/2309.07852/) (2023)
- [Factcheck-Bench: Fine-Grained Evaluation Benchmark for Automatic Fact-checkers](https://mlsanity.com/paper/2311.09000/) (2023)
- [Cognitive Dissonance: Why Do Language Model Outputs Disagree with Internal Representations of Truthfulness?](https://mlsanity.com/paper/2312.03729/) (2023)
- [Towards Automated Factchecking: Developing an Annotation Schema and Benchmark for Consistent Automated Claim Detection](https://mlsanity.com/paper/1809.08193/) (2018)
- [Molecular Facts: Desiderata for Decontextualization in LLM Fact Verification](https://mlsanity.com/paper/2406.20079/) (2024)
- [Beyond Factuality: A Comprehensive Evaluation of Large Language Models as Knowledge Generators](https://mlsanity.com/paper/2310.07289/) (2023)
- [AFaCTA: Assisting the Annotation of Factual Claim Detection with Reliable LLM Annotators](https://mlsanity.com/paper/2402.11073/) (2024)

## Built on by (most cited first, 25 of 25)

- [A Survey of LLM-based Deep Search Agents: Paradigm, Optimization, Evaluation, and Challenges](https://mlsanity.com/paper/2508.05668/) (2025)
- [Recent Advances in Large Langauge Model Benchmarks against Data Contamination: From Static to Dynamic Evaluation](https://mlsanity.com/paper/2502.17521/) (2025)
- [Learning to Reason for Factuality](https://mlsanity.com/paper/2508.05618/) (2025)
- [Hallucination to Truth: A Review of Fact-Checking and Factuality Evaluation in Large Language Models](https://mlsanity.com/paper/2508.03860/) (2025)
- [FaStfact: Faster, Stronger Long-Form Factuality Evaluations in LLMs](https://mlsanity.com/paper/2510.12839/) (2025)
- [FACTORY: A Challenging Human-Verified Prompt Set for Long-Form Factuality](https://mlsanity.com/paper/2508.00109/) (2025)
- [VeriTaS: The First Dynamic Benchmark for Multimodal Automated Fact-Checking](https://mlsanity.com/paper/2601.08611/) (2026)
- [Real-World Gaps in AI Governance Research](https://mlsanity.com/paper/2505.00174/) (2025)
- [PII Jailbreaking in LLMs via Activation Steering Reveals Personal Information Leakage](https://mlsanity.com/paper/2507.02332/) (2025)
- [MAD-Fact: A Multi-Agent Debate Framework for Long-Form Factuality Evaluation in LLMs](https://mlsanity.com/paper/2510.22967/) (2025)
- [FactReasoner: A Probabilistic Approach to Long-Form Factuality Assessment for Large Language Models](https://mlsanity.com/paper/2502.18573/) (2025)
- [Comprehensiveness Metrics for Automatic Evaluation of Factual Recall in Text Generation](https://mlsanity.com/paper/2510.07926/) (2025)
- [A Geometric Analysis of Small-sized Language Model Hallucinations](https://mlsanity.com/paper/2602.14778/) (2026)
- [Two-Level Meta-Rubrics for Evaluating Open-Ended Generation: GAMUT, a Benchmark for Factual Completeness](https://mlsanity.com/paper/2607.19322/) (2026)
- [Even GPT-5.2 Can't Count to Five: The Case for Zero-Error Horizons in Trustworthy LLMs](https://mlsanity.com/paper/2601.15714/) (2026)
- [The Deliberative Illusion: Diagnosing Factual Attrition and Stance Homogenization in Multi-Agent LLM Deliberation](https://mlsanity.com/paper/2606.03032/) (2026)
- [Iterative Tree Analysis for Medical Critics](https://mlsanity.com/paper/2501.10642/) (2025)
- [PerHalluEval: Persian Hallucination Evaluation Benchmark for Large Language Models](https://mlsanity.com/paper/2509.21104/) (2025)
- [Assessing Web Search Credibility and Response Groundedness in Chat Assistants](https://mlsanity.com/paper/2510.13749/) (2025)
- [Not All Needles Are Found: How Fact Distribution and Don't Make It Up Prompts Shape Retrieval, Reasoning, and Hallucination in Long-Context LLMs](https://mlsanity.com/paper/2601.02023/) (2026)
- [Trust but Verify: Introducing DAVinCI -- A Framework for Dual Attribution and Verification in Claim Inference for Language Models](https://mlsanity.com/paper/2604.21193/) (2026)
- [Only Say What You Know: Calibration-Aware Generation for Long-Form Factuality](https://mlsanity.com/paper/2605.01749/) (2026)
- [CommunityFact: A Dynamic, Multilingual, Multi-domain Benchmark for Misinformation Detection in the Wild](https://mlsanity.com/paper/2605.30241/) (2026)
- [SWORD: Wikidata-based Distortions Reveal Hidden Cross-Lingual Inconsistencies in LLM Factual Error Rejection](https://mlsanity.com/paper/2609.09349/) (2026)
- [Do Factuality Methods Make LLMs More Factual? Reproducibility and Fair Comparison](https://mlsanity.com/paper/2610.18415/) (2026)

## Abstract (arXiv)

The rapid adoption of language models (LMs) across diverse applications has raised concerns about their factuality, i.e., their consistency with real-world facts. We first present VERIFY (Verification and Evidence RetrIeval for FactualitY evaluation), a pipeline to evaluate LMs' factuality in real-world user interactions. VERIFY considers the verifiability of LM-generated content and categorizes content units as supported, unsupported, or undecidable based on Web-retrieved evidence. Importantly, factuality judgment by VERIFY correlates better with human evaluations than existing methods. Using VERIFY, we identify "hallucination prompts" across diverse topics, i.e., those eliciting the highest rates of incorrect (unsupported) and inconclusive (undecidable) LM responses. These prompts form FACTBENCH, a dataset of 1K prompts across 150 fine-grained topics. Our dataset captures emerging factuality challenges in real-world LM interactions and can be regularly updated with new prompts. We benchmark widely-used LMs from GPT, Gemini, and Llama families on FACTBENCH, yielding the following key findings: (i) Proprietary models exhibit better factuality, with decreased performance from Easy to Hard hallucination prompts. (ii) Llama3.1-405B-Instruct shows comparable or lower factual precision than Llama3.1-70B-Instruct across all evaluation methods due to its higher subjectivity that leads to more content labeled as undecidable. (iii) Gemini1.5-Pro shows a significantly higher refusal rate, with over-refusal in 25% of cases.