mlsanityData to 14 Oct 2026

Recognizing Is Not Reversing: A Controlled Inversion Test of Fact-Preserving News Framing

arXiv 2609.11769 ↗First posted 10 Sep 2026cs.CL cs.AIcited by 0

Not on the map. Why?

Abstract (arXiv)

Large language models (LLMs) are increasingly used to analyze and rewrite news, yet current framing studies mainly evaluate generation, detection, or whether rewritten text appears more neutral. They do not directly show whether a model can undo a known framing transformation while keeping the facts fixed. We introduce a controlled inversion test over three established textual realizations of framing: evaluative lexis, agency realization, and information salience. Across 60 news articles and three intervention strengths, this yields 540 paired variants with preserved atomic facts and recorded edits. Across Qwen, DeepSeek, and Kimi, factual preservation remains near 0.84, whereas intervention reversal is 0.044--0.068. Even when both framing type and direction are recognized correctly, pooled reversal reaches 0.071. These results reveal a clear separation between factual fidelity, framing recognition, and framing inversion: recognizing how an article is framed does not imply that the framing can be undone.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 21 of its 34 references that are in the map. First posted on arXiv 10 Sep 2026.

Builds on

21 of 34 references.

References · 21

  1. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,459
  2. 2021Fantastically Ordered Prompts and Where to Find Them: Overcoming Few-Shot Prompt Order Sensitivitycited by 1,135
  3. 2023Language Models Don't Always Say What They Think: Unfaithful Explanations in Chain-of-Thought Promptingcited by 1,122
  4. 2023ChatGPT Outperforms Crowd-Workers for Text-Annotation Taskscited by 870
  5. 2023Quantifying Language Models' Sensitivity to Spurious Features in Prompt Design or: How I learned to start worrying about prompt formattingcited by 817
  6. 2022Out of One, Many: Using Language Models to Simulate Human Samplescited by 752
  7. 2020CrowS-Pairs: A Challenge Dataset for Measuring Social Biases in Masked Language Modelscited by 710
  8. 2023Measuring Faithfulness in Chain-of-Thought Reasoningcited by 703
  9. 2021BBQ: A Hand-Built Bias Benchmark for Question Answeringcited by 701
  10. 2023Bias and Fairness in Large Language Models: A Surveycited by 639
  11. 2021BOLD: Dataset and Metrics for Measuring Biases in Open-Ended Language Generationcited by 412
  12. 2023Can Large Language Models Transform Computational Social Science?cited by 250
9 more
  1. 2019In Plain Sight: Media Bias Through the Lens of Factual Reportingcited by 68
  2. 2022Inducing Positive Perspectives with Text Reframingcited by 31
  3. 2021MBIC -- A Media Bias Annotation Dataset Including Annotator Characteristicscited by 24
  4. 2024Larger Language Models Don't Care How You Think: Why Chain-of-Thought Prompting Fails in Subjective Taskscited by 8
  5. 2024Decoding News Narratives: A Critical Analysis of Large Language Models in Framing Detectioncited by 6
  6. 2025Thinking in a Crowd: How Auxiliary Information Shapes LLM Reasoningcited by 5
  7. 2025Frame In, Frame Out: Measuring Framing Bias in LLM-Generated News Summariescited by 2
  8. 2026When Bigger Isn't Better: A Comprehensive Fairness Evaluation of Political Bias in Multi-News Summarisationcited by 2
  9. 2026UnBias-Plus: Detect, Explain, and Rewrite Biascited by 1

Built on by

Not cited yet.