mlsanityData to 7 Oct 2026

Improving Factual Consistency of News Summarization by Contrastive Preference Optimization

arXiv 2310.19347 ↗First posted 30 Oct 2023cs.CL cs.AIcited by 6

On the map: LLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment

Abstract (arXiv)

Despite the recent progress in news summarization made by large language models (LLMs), they often generate summaries that are factually inconsistent with original articles, known as "hallucinations" in text generation. Unlike previous small models (e.g., BART, T5), current LLMs make fewer silly mistakes but more sophisticated ones, such as imposing cause and effect, adding false details, overgeneralizing, etc. These hallucinations are challenging to detect through traditional methods, which poses great challenges for improving the factual consistency of text summarization. In this paper, we propose Contrastive Preference Optimization (CPO) to disentangle the LLMs' propensities to generate faithful and fake content. Furthermore, we adopt a probing-based specific training method to improve their capacity of distinguishing two types of propensities. In this way, LLMs can execute the instructions more accurately and have enhanced perception of hallucinations. Experimental results show that CPO significantly improves the reliability of summarization based on LLMs.

Key facts. Cited by 6 papers in the map (data through 7 Oct 2026). Placed in Human values and AI feedback alignment, one of 10 directions in LLM alignment / RLHF / preference optimisation. Builds on 45 of its 58 references that are in the map. First posted on arXiv 30 Oct 2023.

Builds on

45 of 58 references.

Closest ancestors · 2

Same direction, newest first

  1. 2022On Improving Summarization Factual Consistency from Natural Language Feedback20 Dec
  2. 2020Learning to summarize from human feedback2 Sep

Other references · 43

  1. 2019Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformercited by 14,491
  2. 2017Proximal Policy Optimization Algorithmscited by 14,441
  3. 2023LLaMA: Open and Efficient Foundation Language Modelscited by 13,958
  4. 2023Llama 2: Open Foundation and Fine-Tuned Chat Modelscited by 12,865
  5. 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 7,916
  6. 2019BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehensioncited by 5,874
  7. 2022OPT: Open Pre-trained Transformer Language Modelscited by 3,464
  8. 2021TruthfulQA: Measuring How Models Mimic Human Falsehoodscited by 2,919
  9. 2022Locating and Editing Factual Associations in GPTcited by 2,609
  10. 2022Language Models (Mostly) Know What They Knowcited by 1,783
  11. 2018Don't Give Me the Details, Just the Summary! Topic-Aware Convolutional Neural Networks for Extreme Summarizationcited by 1,308
  12. 2023Inference-Time Intervention: Eliciting Truthful Answers from a Language Modelcited by 1,259
31 more
  1. 2019PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarizationcited by 1,100
  2. 2019BERT Rediscovers the Classical NLP Pipelinecited by 1,046
  3. 2015Teaching Machines to Read and Comprehendcited by 1,039
  4. 2021GLM: General Language Model Pretraining with Autoregressive Blank Infillingcited by 1,026
  5. 2020On Faithfulness and Factuality in Abstractive Summarizationcited by 895
  6. 2023The Flan Collection: Designing Data and Methods for Effective Instruction Tuningcited by 734
  7. 2022Discovering Latent Knowledge in Language Models Without Supervisioncited by 733
  8. 2023OpenAssistant Conversations -- Democratizing Large Language Model Alignmentcited by 712
  9. 2023The Internal State of an LLM Knows When It's Lyingcited by 644
  10. 2019Evaluating the Factual Consistency of Abstractive Text Summarizationcited by 578
  11. 2019Text Summarization with Pretrained Encoderscited by 553
  12. 2023Benchmarking Large Language Models for News Summarizationcited by 444
  13. 2022Emergent World Representations: Exploring a Sequence Model Trained on a Synthetic Taskcited by 417
  14. 2021SummaC: Re-Visiting NLI-based Models for Inconsistency Detection in Summarizationcited by 402
  15. 2020Asking and Answering Questions to Evaluate the Factual Consistency of Summariescited by 376
  16. 2022Self-critiquing models for assisting human evaluatorscited by 372
  17. 2020FEQA: A Question Answering Evaluation Framework for Faithfulness Assessment in Abstractive Summarizationcited by 292
  18. 2022Relative representations enable zero-shot latent space communicationcited by 175
  19. 2023Verify-and-Edit: A Knowledge-Enhanced Chain-of-Thought Frameworkcited by 139
  20. 2023Human-like Summarization Evaluation with ChatGPTcited by 126
  21. 2021CLIFF: Contrastive Learning for Improving Faithfulness and Factuality in Abstractive Summarizationcited by 120
  22. 2019Don't Say That! Making Inconsistent Dialogue Unlikely with Unlikelihood Trainingcited by 110
  23. 2022DisentQA: Disentangling Parametric and Contextual Knowledge with Counterfactual Question Answeringcited by 93
  24. 2023TrueTeacher: Learning Factual Consistency Evaluation with Large Language Modelscited by 82
  25. 2023Element-aware Summarization with Large Language Models: Expert-aligned Evaluation and Chain-of-Thought Methodcited by 76
  26. 2023Factually Consistent Summarization via Reinforcement Learning with Textual Entailment Feedbackcited by 71
  27. 2020Multi-Fact Correction in Abstractive Text Summarizationcited by 66
  28. 2023Large Language Models are Diverse Role-Players for Summarization Evaluationcited by 62
  29. 2022FactPEGASUS: Factuality-Aware Pre-training and Fine-tuning for Abstractive Summarizationcited by 49
  30. 2022Contrastive Learning Reduces Hallucination in Conversationscited by 39
  31. 2023Calibrating Likelihoods towards Consistency in Summarization Modelscited by 4

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 2; 2024: 0; 2025: 1; 2026: 3.

6 papers.

  1. 2023A Survey on Hallucination in Large Language Models: Principles, Taxonomy, Challenges, and Open Questionscited by 1,582
  2. 2023Adapting Large Language Models for Education: Foundational Capabilities, Potentials, and Challengescited by 13
  3. 2025ReFeed: Multi-dimensional Summarization Refinement with Reflective Reasoning on Feedbackcited by 4
  4. 2026BLooP: Zero-Shot Abstractive Summarization using Large Language Models with Bigram Lookahead Promotioncited by 0
  5. 2026LLMs Should Incorporate Explicit Mechanisms for Human Empathycited by 0
  6. 2026Beyond Overlap Metrics: Rewarding Reasoning and Preferences for Faithful Multi-Role Dialogue Summarizationcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026Optimising Factual Consistency in Summarisation via Preference Learning from Multiple Imperfect Metrics
  2. 2025Learning from Self Critique and Refinement for Faithful LLM Summarization
  3. 2023Evaluating Factual Consistency of Summaries with Large Language Models
  4. 2024Model-based Preference Optimization in Abstractive Summarization without Human Feedback
  5. 2025Multi-round, Chain-of-thought Post-editing for Unfaithful Summaries
  6. 2026Hallucination Detection-Guided Preference Optimization for Clinical Summarization
  7. 2024Improving Faithfulness of Large Language Models in Summarization via Sliding Generation and Self-Consistency
  8. 2023Fine-tuning Language Models for Factuality