CS Paper Radar

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

Yi Liu

cs.CL cs.AI ·submitted 2026-09-10 16:23 UTC · v1

Abstract

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.

Closest papers

Nearest neighbours by meaning across the whole index.

Framing-Induced Vulnerability in Large Language Models: Evidence from Chinese News Headlines

cs.CL ·2026-11-10 ·Mo Zhou, Yiwei Zhang
similarity 0.68
Abstract · 7 pages, 5 figures

Large language models increasingly evaluate, route, and moderate news, so a model's judgment of a headline has become a gatekeeping decision. We ask whether that judgment follows the fact being reported or the way the fact is framed. We combine an ecological audit of 63,926 Chinese news headlines with a controlled experiment. The experiment covers 30 real events, 30 synthetic events, and 29 natural headline pairs, tested on four Chinese models. Changing only the headline frame shifted both perceived misleadingness and event severity, and the direction and ordering of the shifts were the same for DeepSeek, MiMo, Dots, and LongCat. Emotionally intensified headlines raised severity ratings by 0.65 to 0.88 points, whereas positively reframed headlines lowered them by 0.34 to 0.52 points. Supplying the article body reduced this severity bias by 54 to 80 percent, and all four reductions were significant at p < .01. A small human reference study showed the same direction of effect. Headline-only LLM evaluation therefore carries a measurable robustness gap, and framing sensitivity belongs among the reliability properties that deployment and governance check.

0.14

What Breaks When Facts Become Stories? The Distortions Source-Agreement Metrics Miss When Models Retell Real Events

cs.AI ·2026-10-29 ·Kwangsuk Park, JiWoong Yang, Vaniushkina Kseniia
publishedcode/data similarity 0.67
Abstract · 14 pages, 7 figures, 11 tables. Accepted at the Trust-AI-Eval (TAE) Workshop at NeurIPS 2026. Data and code: https://github.com/Retold2K/ret…

Language models are now widely used to turn recorded facts into media for public consumption, yet evaluation of such output still assumes agreement with the source as its measure of success. In reconstruction, where departing from the source can itself be the goal, this criterion measures the wrong construct: applied as is, it counts intended departures as errors; withheld, it lets real distortion through. We had three model families rewrite 2,000 news articles as summaries and as stories, measured what changed on every axis of the five Ws and how, and asked of each detected distortion whether the assigned perspective warranted it, which set a third of them aside. Time and manner were not lost but replaced, while place and causation lost the expressions that carry them. Accountability distortion behaved otherwise. Barely present in summaries, it rose up to sevenfold in stories, and of those whose direction could be determined, two in three lightened the perpetrator's responsibility. Such distortions survive claim-level fact checking and leave the retelling grammatically cleaner than its source. Where our judges disagreed, three authors relabeling the same cases converged on whether a distortion had occurred but not on whether it was warranted, a boundary contested rather than merely hard to detect. To support work on detecting and reducing them, we release ReTold-2k, a reconstructable corpus with the scoring protocol and per-judge baselines kept unaggregated.(https://github.com/Retold2K/retold2k).

0.42

The Truth Beneath the Tokens: Prompt-Space Recovery of Knowledge Edits in Language Models

cs.CL cs.AI ·2026-11-07 ·Manit Baser, Rachel Yeo Hui Min, Aditya Nawal, Dinil Mon Divakaran, Mohan Gurusamy
similarity 0.65
Abstract

Model editing provides an efficient way to update factual knowledge in transformer models. Recent approaches reverse edits by modifying, masking, or decomposing model weights, but make recovery dependent on weight-space interventions. We instead show that pre-edit knowledge can be recovered through input-space interventions alone, without modifying the edited model's weights. We introduce Memento, a prompt-space mechanism for auditing residual knowledge that learns a reusable recovery signal to steer an edited model toward its pre-edit response. Memento supports both continuous recovery prefixes and discrete textual prompts for standard text interfaces. Across three LLMs and five editing techniques, Memento recovers 80.07% of held-out edited facts, achieving a 9.58% relative improvement over baselines, suggesting that editing may suppress the facts without fully eliminating their recoverability. We further investigate whether prompt-space recovery extends beyond model editing. On pristine, unedited LLMs, Memento improves average factual recall by 53.9% relative to standard inference. These results demonstrate the potential of prompt-space recovery for auditing residual knowledge in edited models and investigating the gap between factual knowledge and its accessibility during inference. We will release our source code.

0.22

Framing by Wording, Framing by Selection: A Large-Scale Two-Dimensional Audit of French News Headlines, 2022-2025

cs.CL ·2026-08-12 ·Amr Sobhy
published similarity 0.64
Abstract · 20 pages, 1 figure, includes appendices. Accepted for oral presentation at ICNLSP 2026

News headlines frame public issues both by what they select and by how they word it, yet computational framing work typically collapses these operations into a single score. We introduce a two-dimensional framework that separates salience framing, measured through four wording devices (loaded vocabulary, blame attribution, threat framing, rhetorical question), from selection framing, measured through outlet-level story-form and high-charge distributions. We build a 10,000-headline French supervision set using three LLM annotators with majority-vote resolution and human arbitration, validate the labels against two annotator-independent blind human studies, and apply the strongest classifier to 902,111 deduplicated headlines from 25 French outlets (2022-2025). Three main findings emerge. First, salience and selection divergence are positively correlated yet leave nearly half of outlet-level variance unexplained, populating interpretively distinct off-diagonal cells in a four-cell outlet typology. Second, default classification thresholds systematically inflate corpus-level salience estimates; a precision-floor recalibration protocol corrects this distortion. Third, group-mention analysis reveals sharply unequal salience contexts: headlines mentioning Jews, the Far-right, and Muslims carry the highest detected salience rates, which broad event-context composition does not fully explain (residuals are descriptive, not same-event causal estimates; per-group lexicon precision is reported alongside). To our knowledge, this is the largest framing-focused French headline audit to date; we release the supervision set, lexicons, and analysis code.

0.25

Where Does Rhetorical Distortion Live? A Multi-Platform, Multi-Language Measurement of Influencer Content

cs.SI cs.CLcs.CY ·2026-10-19 ·Xucheng Yu, Jiangping Chen, Haohan Wang
similarity 0.63
Abstract · 12 pages

Much of the manipulation in influencer content is not false but disproportionate, framing accurate facts as historic, urgent, or catastrophic. Existing approaches either verify factual claims or detect clickbait and propaganda in single-platform, single-language text; however, they overlook where such framing originates, because they treat a post and the replies beneath it as one unit. To address this gap, we measure rhetorical distortion with one instrument applied identically across six platforms and two languages, classifying creator posts and their replies separately. Our three-tier pipeline detects five literature-grounded dimensions (Significance Inflation, Anxiety Manufacturing, Novelty Claims, Loaded Language, and Temporal Distortion): bilingual lexical rules propose candidates, negative filters cancel known false positives, and an LLM verifies or relabels the candidates. This design keeps labels transparent and comparable across platforms, calls the LLM on only 14% of posts, and supports validation against human annotators and a second model. On 13,060 posts from 330 accounts on RSS/newsletters, YouTube, Bluesky, Reddit, Weibo, and X/Twitter, platform is far more strongly associated with account-level distortion than content category (partial $η^2=0.49$ vs. $0.09$), and X/Twitter scores highest (DI 7.74 vs. 0.78 for RSS). On X/Twitter, the top-5 replies under a post carry 3.5x the distortion of the tweet they answer (1.5x at equal length), while creator tweets alone are indistinguishable from YouTube and Bluesky; two human annotators and GPT-4o confirm this gap. Measured distortion on threaded platforms thus lives largely in the audience layer, which calls for measuring creator and reply text separately.

0.24

When Steering Shifts Discourse: Identifying and Intervening on Discourse-Level Changes due to Activation Steering in LLMs

cs.CL cs.AI ·2026-10-30 ·Shivam Adarsh, Hetianyu Huang, Lau Safin, Maria Maistro, Christina Lioma
similarity 0.63
Abstract

Activation steering is an inference-time technique used to control or modify the behavior, tone, or safety properties of a large language model (LLM) without retraining the model's weights or editing its input prompt, but simply by adding a vector to its residual-stream activations. However, activation steering also changes the generated text in unintended ways, unrelated to intended behavior change, that are not well understood. Prior work has measured these changes only through coarse signals such as perplexity. In this work, we show that steering systematically produces significant changes in discourse dimensions of the generated text, such as its formality, use of evidence, and ambiguity. We present a way of countering this by adding a second, complementary vector that targets the affected discourse dimension. This intervention restores about 78% of the unintended discourse changes while preserving the intended steering behavior in about 93% of the generations.

0.20

Who Judges the Frame? Auditing Multimodal LLM Judges for News Framing Across Event-Level Perspectives

cs.CY ·2026-09-04 ·Antonela Tommasel, Markus Schedl
published similarity 0.62
Abstract

News coverage of major world events is shaped not only by what is reported, but also by how events are framed through text, images and their combination. At the same time, Large Language Models (LLMs), including multimodal LLMs, are increasingly used as scalable instruments for analysing framing, sentiment, ideological slant and perspective differences in multimodal media datasets. This creates a methodological challenge. When used as measurement instruments, LLM outputs may reflect not only content properties, but also model-specific tendencies, prompt design choices, and social, political, cultural, linguistic or modality-specific assumptions. This work audits LLMs as instruments for large-scale framing and perspective analysis in multimodal news coverage. Using an event-centered dataset of 2025--2026 news coverage, where each event includes left-, center- and right-oriented articles about the same headline, we combine embedding-based measures of within-event viewpoint similarity with model-based assessments of framing constructs across modality-specific and metadata-visible input conditions. Rather than treating either dataset labels or model outputs as ground truth, our goal is to examine the usefulness and limitations of LLM-based media analysis. The study contributes an audit protocol that highlights the need to report modality effects, metadata sensitivity, and prompt-induced artifacts alongside substantive claims about news framing and ideological viewpoint differences.

0.30

Improving Atomic-Fact Recall via Focused Views in Unstructured Knowledge Editing

cs.CL cs.AI ·2026-10-02 ·Ding Wu, Ye Zhang, Haoyu Wang, Tianci Liu
similarity 0.62
Abstract · The first two authors contributed equally

Large language models (LLMs) increasingly serve as general-purpose interfaces to factual knowledge, but their parameters do not automatically reflect information that changes after pretraining. Knowledge editing (KE) provides a targeted alternative to costly retraining by modifying selected knowledge and preserving unrelated knowledge and general capabilities. Conventional KE uses structured factual triples, whereas unstructured KE (UKE) uses free-form passages containing multiple facts. Nonetheless, existing UKE editors exhibit a failure mode known as context reliance: edited LLMs can often reproduce the editing passage but fail to reliably recall its individual facts without the original passage context. We identify context-induced difficulty underestimation under the standard passage-level editing objective: later facts receive increasingly rich ground-truth context and consequently incur lower initial losses, making them appear easier to learn. In response, we propose FOVEATED, a plug-and-play framework that constructs focused views of each sentence by randomly shifting the Rotary Position Embedding (RoPE) positions assigned to the keys of its preceding context. The perturbation is applied during editing and removed afterward, leaving the model's native positional encoding unchanged at inference time. We instantiate FOVEATED for both direct-optimization and locate-then-edit editors. We theoretically analyze how FOVEATED counteracts context-induced difficulty underestimation and empirically demonstrate consistent improvements across five KE editors, two LLM backbones, and three benchmarks.

0.16