mlsanityData to 7 Oct 2026

BeliefBank: Adding Memory to a Pre-Trained Language Model for a Systematic Notion of Belief

arXiv 2109.14723 ↗First posted 29 Sep 2021cs.CLcited by 56

Not on the map. Why?

Abstract (arXiv)

Although pretrained language models (PTLMs) contain significant amounts of world knowledge, they can still produce inconsistent answers to questions when probed, even after specialized training. As a result, it can be hard to identify what the model actually "believes" about the world, making it susceptible to inconsistent behavior and simple errors. Our goal is to reduce these problems. Our approach is to embed a PTLM in a broader system that also includes an evolving, symbolic memory of beliefs -- a BeliefBank -- that records but then may modify the raw PTLM answers. We describe two mechanisms to improve belief consistency in the overall system. First, a reasoning component -- a weighted MaxSAT solver -- revises beliefs that significantly clash with others. Second, a feedback component issues future queries to the PTLM using known beliefs as context. We show that, in a controlled experimental setting, these two mechanisms result in more consistent beliefs in the overall system, improving both the accuracy and consistency of its answers over time. This is significant as it is a first step towards PTLM-based architectures with a systematic notion of belief, enabling them to construct a more coherent picture of the world, and improve over time without model retraining.

Key facts. Cited by 56 papers in the map (data through 7 Oct 2026). Builds on 21 of its 43 references that are in the map. First posted on arXiv 29 Sep 2021.

Builds on

21 of 43 references.

References · 21

  1. 2020Retrieval-Augmented Generation for Knowledge-Intensive NLP Taskscited by 9,160
  2. 2020REALM: Retrieval-Augmented Language Model Pre-Trainingcited by 2,126
  3. 2019Language Models as Knowledge Bases?cited by 1,922
  4. 2018Continual Lifelong Learning with Neural Networks: A Reviewcited by 1,313
  5. 2021What Makes Good In-Context Examples for GPT-3?cited by 1,184
  6. 2016ConceptNet 5.5: An Open Multilingual Graph of General Knowledgecited by 1,065
  7. 2020How Much Knowledge Can You Pack Into the Parameters of a Language Model?cited by 713
  8. 2021Measuring and Improving Consistency in Pretrained Language Modelscited by 361
  9. 2019What BERT is not: Lessons from a new suite of psycholinguistic diagnostics for language modelscited by 317
  10. 2020Transformers as Soft Reasoners over Languagecited by 265
  11. 2019Negated and Misprimed Probes for Pretrained Language Models: Birds Can Talk, But Cannot Flycited by 240
  12. 2019Commonsense Knowledge Mining from Pretrained Modelscited by 189
9 more
  1. 2020Unsupervised Commonsense Question Answering with Self-Talkcited by 161
  2. 2020Logic-Guided Data Augmentation and Regularization for Consistent Question Answeringcited by 77
  3. 2019A Logic-Driven Framework for Consistency of Neural Modelscited by 64
  4. 2016Tracking the World State with Recurrent Entity Networkscited by 51
  5. 2020Conditionally Adaptive Multi-Task Learning: Improving Transfer Learning in NLP Using Fewer Parameters & Less Datacited by 50
  6. 2018Adversarially Regularising Neural NLI Models to Integrate Logical Background Knowledgecited by 45
  7. 2021General-Purpose Question-Answering with Macawcited by 40
  8. 2020Remembering for the Right Reasons: Explanations Reduce Catastrophic Forgettingcited by 33
  9. 2020Obtaining Faithful Interpretations from Compositional Neural Networkscited by 23

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 2; 2022: 13; 2023: 13; 2024: 15; 2025: 4; 2026: 9.

56 papers.

Sort:

By area:

  1. 2022Memory-Based Model Editing at Scalecited by 443
  2. 2023Adaptive Chameleon or Stubborn Sloth: Revealing the Behavior of Large Language Models in Knowledge Conflictscited by 333
  3. 2023Evaluating the Ripple Effects of Knowledge Editing in Language Modelscited by 238
  4. 2024Do Large Language Models Latently Perform Multi-Hop Reasoning?cited by 165
  5. 2022Maieutic Prompting: Logically Consistent Reasoning with Recursive Explanationscited by 159
  6. 2022A Review on Language Models as Knowledge Basescited by 156
  7. 2022Towards Faithful Model Explanation in NLP: A Surveycited by 129
  8. 2021Do Language Models Have Beliefs? Methods for Detecting, Updating, and Visualizing Model Beliefscited by 79
  9. 2022Measuring Causal Effects of Data Statistics on Language Model's `Factual' Predictionscited by 57
  10. 2025Empowering LLMs with Logical Reasoning: A Comprehensive Surveycited by 51
  11. 2022Enhancing Self-Consistency and Performance of Pre-Trained Language Models through Natural Language Inferencecited by 40
  12. 2022Entailer: Answering Questions with Faithful and Truthful Chains of Reasoningcited by 38
38 more
  1. 2022Towards Teachable Reasoning Systems: Using a Dynamic Memory of User Feedback for Continual System Improvementcited by 37
  2. 2023Two Failures of Self-Consistency in the Multi-Step Reasoning of LLMscited by 36
  3. 2021Learning to Repair: Repairing model output errors after deployment using a dynamic memory of feedbackcited by 35
  4. 2023How Do Large Language Models Capture the Ever-changing World Knowledge? A Review of Recent Advancescited by 33
  5. 2023Say What You Mean! Large Language Models Speak Too Positively about Negative Commonsense Knowledgecited by 32
  6. 2024Logically Consistent Language Models via Neuro-Symbolic Integrationcited by 30
  7. 2024Belief Revision: The Adaptability of Large Language Models Reasoningcited by 26
  8. 2022Factual Consistency of Multilingual Pretrained Language Modelscited by 15
  9. 2023Separating form and meaning: Using self-consistency to quantify task understanding across multiple sensescited by 15
  10. 2023Language Models with Rationalitycited by 15
  11. 2023Semantic Consistency for Assuring Reliability of Large Language Modelscited by 15
  12. 2022Do language models have coherent mental models of everyday things?cited by 13
  13. 2025Improving Consistency in Large Language Models through Chain of Guidancecited by 13
  14. 2024Enabling LLM Knowledge Analysis via Extensive Materializationcited by 12
  15. 2024Investigating Context-Faithfulness in Large Language Models: The Roles of Memory Strength and Evidence Stylecited by 11
  16. 2023GLUECons: A Generic Benchmark for Learning Under Constraintscited by 9
  17. 2024Desiderata for the Context Use of Question Answering Systemscited by 7
  18. 2024AI, Meet Human: Learning Paradigms for Hybrid Decision Making Systemscited by 7
  19. 2024From Form(s) to Meaning: Probing the Semantic Depths of Language Models Using Multisense Consistencycited by 7
  20. 2024Defining Knowledge: Bridging Epistemology and Large Language Modelscited by 7
  21. 2022Breakpoint Transformers for Modeling and Tracking Intermediate Beliefscited by 6
  22. 2023Prompt to be Consistent is Better than Self-Consistent? Few-Shot and Zero-Shot Fact Verification with Pre-trained Language Modelscited by 6
  23. 2024Hire Me or Not? Examining Language Model's Behavior with Occupation Attributescited by 5
  24. 2022Norm Participation Grounds Languagecited by 4
  25. 2024Evaluating Biases in Context-Dependent Health Questionscited by 4
  26. 2025AI Debaters are More Persuasive when Arguing in Alignment with Their Own Beliefscited by 4
  27. 2025ESCA: Contextualizing Embodied Agents via Scene-Graph Generationcited by 4
  28. 2023Modularity in Deep Learning: A Surveycited by 3
  29. 2024Untangle the KNOT: Interweaving Conflicting Knowledge and Reasoning Skills in Large Language Modelscited by 3
  30. 2026Dutch Books for Language Modelscited by 3
  31. 2022Memory-assisted prompt editing to improve GPT-3 after deploymentcited by 2
  32. 2023Exposing and Addressing Cross-Task Inconsistency in Unified Vision-Language Modelscited by 2
  33. 2024Does Instruction Tuning Make LLMs More Consistent?cited by 2
  34. 2023Neural Conversation Models and How to Rein Them in: A Survey of Failures and Fixescited by 1
  35. 2024From Models to Microtheories: Distilling a Model's Topical Knowledge for Grounded Question Answeringcited by 1
  36. 2026An Agentic Framework for Neuro-Symbolic Programmingcited by 1
  37. 2026Beyond Calibration: Do a Typed-Decision Model's Probabilities Obey the Probability Axioms?cited by 1
  38. 2024FairBelief -- Assessing Harmful Beliefs in Language Modelscited by 0

50 most cited of 56 All 56 (JSON)