AI Paper Atlas

Research map / Label-efficient and robust learning / OOD and anomaly detection

Bot Meets Shortcut: How Can LLMs Aid in Handling Unknown Invariance OOD Scenarios?

Shiyan Zheng, Herun Wan, Minnan Luo, Junhang Huang

AAAI 2026 · Label-efficient and robust learning › OOD and anomaly detection · approach: Robust fine-tuning for OOD generalization · 1 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
train on shortcut-correlated sentiment data instead of standard data (ablation)RoBERTa+MLP · Twibot-20 shortcut testaccuracy: .682 -> .051Hurt
train on shortcut-correlated sentiment data instead of standard data (ablation)RoBERTa+MLP · Cresci-2017 shortcut testaccuracy: .884 -> .287Hurt
train on shortcut-correlated topic data instead of standard data (ablation)RoBERTa+MLP · Cresci-2015 shortcut testaccuracy: .985 -> .941No clear effect
add AMR+CIGA causal debiasingRoBERTa shortcut-trained · Cresci-2017 shortcut test (sentiment)accuracy: .287 -> .271No clear effect
add AMR+CIGA causal debiasingRoBERTa shortcut-trained · Cresci-2015 shortcut test (topic)accuracy: .941 -> .774Hurt
add LLM counterfactual Text-Level debiasingRoBERTa shortcut-trained · Twibot-20 shortcut test (sentiment)accuracy: .051 -> .229Improved
add Dataset-Level counterfactual debiasingRoBERTa shortcut-trained · Twibot-20 shortcut test (topic)accuracy: .164 -> .607Improved
add Model-Level contrastive fine-tuning of the feature extractorRoBERTa shortcut-trained · Twibot-20 shortcut test (sentiment)accuracy: .051 -> .415Improved
add Model-Level debiasingRoBERTa shortcut-trained · Cresci-2017 shortcut test (topic)accuracy: .587 -> .839Improved
Dataset-Level and Model-Level debiasing on values shortcutRoBERTa shortcut-trained · Cresci-2015 shortcut test (values)accuracy: .659 -> .633 (dataset), .615 (model); but Text-Level .762Mixed
all three LLM-based mitigation levels on text-based modelsRoBERTa-based detectors · shortcut scenariosrelative accuracy: average 59% relative improvement (text), 53% (graph)Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy