#TruthOverTricks
Truth over Tricks: Measuring and Mitigating Shortcut Learning in Misinformation Detection
Misinformation detection models often rely on superficial cues (i.e., \emph{shortcuts}) that correlate with misinformation in training data but fail to generalize to the diverse and evolving nature of real-world misinformation. This issue is exacerbated by large language models (LLMs), which can easily generate convincing misinformation through simple prompts. We introduce TruthOverTricks, a unified evaluation paradigm for measuring shortcut learning in misinformation detection. TruthOverTricks categorizes shortcut behaviors into intrinsic shortcut induction and extrinsic shortcut injection, and evaluates seven representative detectors across 14 popular benchmarks, along with two new factual misinformation datasets, NQ-Misinfo and Streaming-Misinfo. Empirical results reveal that existing detectors suffer severe performance degradation when exposed to both naturally occurring and adversarially crafted shortcuts. To address this, we propose SMF, an LLM-augmented data augmentation framework that mitigates shortcut reliance through paraphrasing, factual summarization, and sentiment normalization. SMF consistently enhances robustness across 16 benchmarks, encouraging models to rely on deeper semantic understanding rather than shortcut cues. To promote the development of misinformation detectors, we have published the resources publicly at https://github.com/whr000001/TruthOverTricks.
arxiv.org
June 5, 2025 at 2:45 AM
Entertaining a fallacy in debate is like allowing poison at a feast — it does not matter how elegant the banquet, the end is still death.

• #LogicMatters
• #CriticalThinking
• #ReasonOverRhetoric
• #TruthOverTricks
• #FallacyFree
• #PhilosophyMatters
• #Reasoning
#ThinkDeeply
#DebateEthics
June 8, 2025 at 6:35 AM
Seeing JD Vance as a guest = don’t watch.
He does nothing but lie and I’m done giving my time to people who treat politics like performance instead of truth.
#IntegrityMatters #TruthOverTricks #MediaAccountability
October 12, 2025 at 6:46 PM
Don’t get hung up on the decoy. Find what they don’t want you to see!

#TalkOutDaily #DecoyTactics #TruthOverTricks
February 3, 2025 at 12:24 PM
on deeper semantic understanding rather than shortcut cues. To promote the development of misinformation detectors, we have published the resources publicly at https://github.com/whr000001/TruthOverTricks. [6/6 of https://arxiv.org/abs/2506.02350v1]
June 4, 2025 at 6:00 AM
learning in misinformation detection. TruthOverTricks categorizes shortcut behaviors into intrinsic shortcut induction and extrinsic shortcut injection, and evaluates seven representative detectors across 14 popular benchmarks, along with two new [3/6 of https://arxiv.org/abs/2506.02350v1]
June 4, 2025 at 6:00 AM
real-world misinformation. This issue is exacerbated by large language models (LLMs), which can easily generate convincing misinformation through simple prompts. We introduce TruthOverTricks, a unified evaluation paradigm for measuring shortcut [2/6 of https://arxiv.org/abs/2506.02350v1]
June 4, 2025 at 5:59 AM
I was just told that I have to remove “climate” from the title of an ongoing grant if I want to keep it. And publications from that grant cannot include “climate” and other forbidden words. I can’t believe I’m writing this from the United States of America. #AcademicSky
Just received an email saying that it's not advisable to include the following words in future applications for research funding to the US DOT. Here we are.
February 11, 2025 at 7:33 AM
And to attract some people to this cause, I would never use these in the posts I'm suggesting

#FightDisinformation #ExposeLies #ResistPropaganda #DefendTruth #StopTheLies #TruthOverTricks #DisruptDisinfo #ChallengeFalsehoods #ThinkCritically
March 5, 2025 at 7:33 PM