mlsanityData to 14 Oct 2026

ComparisonQA: Evaluating Factuality Robustness of LLMs Through Knowledge Frequency Control and Uncertainty

arXiv 2412.20251 ↗First posted 28 Dec 2024cs.CLcited by 10

Not on the map. Why?

Abstract (arXiv)

The rapid development of LLMs has sparked extensive research into their factual knowledge. Current works find that LLMs fall short on questions around low-frequency entities. However, such proofs are unreliable since the questions can differ not only in entity frequency but also in difficulty themselves. So we introduce ComparisonQA benchmark, containing 283K abstract questions, each instantiated by a pair of high-frequency and low-frequency entities. It ensures a controllable comparison to study the role of knowledge frequency in the performance of LLMs. Because the difference between such a pair is only the entity with different frequencies. In addition, we use both correctness and uncertainty to develop a two-round method to evaluate LLMs' knowledge robustness. It aims to avoid possible semantic shortcuts which is a serious problem of current QA study. Experiments reveal that LLMs, including GPT-4o, exhibit particularly low robustness regarding low-frequency knowledge. Besides, we find that uncertainty can be used to effectively identify high-quality and shortcut-free questions while maintaining the data size. Based on this, we propose an automatic method to select such questions to form a subset called ComparisonQA-Hard, containing only hard low-frequency questions.

Key facts. Cited by 10 papers in the map (data through 14 Oct 2026). Builds on 36 of its 54 references that are in the map. First posted on arXiv 28 Dec 2024.

Builds on

36 of 54 references.

References · 36

  1. 2020Language Models are Few-Shot Learnerscited by 32,419
  2. 2023Llama 2: Open Foundation and Fine-Tuned Chat Modelscited by 13,028
  3. 2023Judging LLM-as-a-Judge with MT-Bench and Chatbot Arenacited by 8,626
  4. 2020Measuring Massive Multitask Language Understandingcited by 7,963
  5. 2020Shortcut Learning in Deep Neural Networkscited by 1,969
  6. 2024Phi-3 Technical Report: A Highly Capable Language Model Locally on Your Phonecited by 1,917
  7. 2024MMLU-Pro: A More Robust and Challenging Multi-Task Language Understanding Benchmarkcited by 1,848
  8. 2023FActScore: Fine-grained Atomic Evaluation of Factual Precision in Long Form Text Generationcited by 1,282
  9. 2022When Not to Trust Language Models: Investigating Effectiveness of Parametric and Non-Parametric Memoriescited by 1,128
  10. 2024Gemma: Open Models Based on Gemini Research and Technologycited by 916
  11. 2021Did Aristotle Use a Laptop? A Question Answering Benchmark with Implicit Reasoning Strategiescited by 743
  12. 2023Is ChatGPT a General-Purpose Natural Language Processing Task Solver?cited by 488
24 more
  1. 2023Large Language Models Are Not Robust Multiple Choice Selectorscited by 486
  2. 2022Large Language Models Struggle to Learn Long-Tail Knowledgecited by 458
  3. 2023FreshLLMs: Refreshing Large Language Models with Search Engine Augmentationcited by 318
  4. 2023How Language Model Hallucinations Can Snowballcited by 311
  5. 2023FacTool: Factuality Detection in Generative AI -- A Tool Augmented Framework for Multi-Task and Multi-Domain Scenarioscited by 251
  6. 2023Do Large Language Models Know What They Don't Know?cited by 243
  7. 2023Head-to-Tail: How Knowledgeable are Large Language Models (LLMs)? A.K.A. Will LLMs Replace Knowledge Graphs?cited by 168
  8. 2023Physics of Language Models: Part 3.2, Knowledge Manipulationcited by 143
  9. 2024Benchmarking Uncertainty Quantification Methods for Large Language Models with LM-Polygraphcited by 105
  10. 2023Factcheck-Bench: Fine-Grained Evaluation Benchmark for Automatic Fact-checkerscited by 87
  11. 2023KCTS: Knowledge-Constrained Tree Search Decoding with Token-Level Hallucination Detectioncited by 45
  12. 2023Vera: A General-Purpose Plausibility Estimation Model for Commonsense Statementscited by 44
  13. 2023COLA: Contextualized Commonsense Causal Reasoning from the Causal Inference Perspectivecited by 22
  14. 2024Conceptual and Unbiased Reasoning in Language Modelscited by 19
  15. 2023AbsPyramid: Benchmarking the Abstraction Ability of Language Models with a Unified Entailment Graphcited by 17
  16. 2024AbsInstruct: Eliciting Abstraction Ability from LLMs through Explanation Tuning with Plausibility Estimationcited by 14
  17. 2024Abstraction-of-Thought Makes Language Models Better Reasonerscited by 11
  18. 2022SubeventWriter: Iterative Sub-event Sequence Generation with Coherence Controllercited by 10
  19. 2024CLR-Fact: Evaluating the Complex Logical Reasoning Capability of Large Language Models over Factual Knowledgecited by 9
  20. 2024Automatic Question-Answer Generation for Long-Tail Knowledgecited by 8
  21. 2023Knowledge Base Completion for Long-Tail Entitiescited by 7
  22. 2024ConstraintChecker: A Plugin for Large Language Models to Reason on Commonsense Knowledge Basescited by 6
  23. 2024Concept-Reversed Winograd Schema Challenge: Evaluating and Improving Robust Reasoning in Large Language Models via Abstractioncited by 4
  24. 2023TILFA: A Unified Framework for Text, Image, and Layout Fusion in Argument Miningcited by 3

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 8; 2026: 2.

10 papers.

  1. 2025Can Indirect Prompt Injection Attacks Be Detected and Removed?cited by 52
  2. 2025CostBench: Evaluating Multi-Turn Cost-Optimal Planning and Adaptation in Dynamic Environments for LLM Tool-Use Agentscited by 33
  3. 2025The Curse of CoT: On the Limitations of Chain-of-Thought in In-Context Learningcited by 30
  4. 2025Revisiting Epistemic Markers in Confidence Estimation: Can Markers Accurately Reflect Large Language Models' Uncertainty?cited by 16
  5. 2025LogiDynamics: Unraveling the Dynamics of Inductive, Abductive and Deductive Logical Inferences in LLM Reasoningcited by 12
  6. 2025Climplicit: Climatic Implicit Embeddings for Global Ecological Taskscited by 10
  7. 2026NOVA: NOise-aware Verbal Confidence CAlibration for Robust Large Language Models in RAG Systemscited by 6
  8. 2025CritiCal: Can Critique Help LLM Uncertainty or Confidence Calibration?cited by 4
  9. 2025From Confidence to Collapse in LLM Factual Robustnesscited by 2
  10. 2026LUMOS: Tracing Parametric Knowledge from Training Data to Behavioral Outputs in LLMscited by 0