Buchumschlag

Mitigating Shortcut Reasoning in Language Models: A Gradient-Aware Training Approach

Fuente: arXiv

Gespeichert in:

Bibliographische Detailangaben
Hauptverfasser:	Cao, Hongyu, Liu, Kunpeng, Wang, Dongjie, Fu, Yanjie
Format:	Preprint
Veröffentlicht:	2026
Schlagworte:	Computation and Language Artificial Intelligence
Online-Zugang:	Acceder al recurso
Tags:	Tag hinzufügen Keine Tags, Fügen Sie den ersten Tag hinzu!

Ähnliche Einträge

Efficient Post-Training Refinement of Latent Reasoning in Large Language Models
von: Wang, Xinyuan, et al.
Veröffentlicht: (2025)

Break the Chain: Large Language Models Can be Shortcut Reasoners
von: Ding, Mengru, et al.
Veröffentlicht: (2024)

Evolving Demonstration Optimization for Chain-of-Thought Feature Transformation
von: Wang, Xinyuan, et al.
Veröffentlicht: (2026)

CURE: Controlled Unlearning for Robust Embeddings -- Mitigating Conceptual Shortcuts in Pre-Trained Language Models
von: Kocak, Aysenur, et al.
Veröffentlicht: (2025)

Deceptive Semantic Shortcuts on Reasoning Chains: How Far Can Models Go without Hallucination?
von: Li, Bangzheng, et al.
Veröffentlicht: (2023)

Mitigating Overthinking in Large Reasoning Language Models via Reasoning Path Deviation Monitoring
von: Guan, Weixin, et al.
Veröffentlicht: (2026)

Spurious Correlations and Beyond: Understanding and Mitigating Shortcut Learning in SDOH Extraction with Large Language Models
von: Sakib, Fardin Ahsan, et al.
Veröffentlicht: (2025)

Revealing and Mitigating the Local Pattern Shortcuts of Mamba
von: You, Wangjie, et al.
Veröffentlicht: (2024)

When Models Outthink Their Safety: Unveiling and Mitigating Self-Jailbreak in Large Reasoning Models
von: Mao, Yingzhi, et al.
Veröffentlicht: (2025)

The Confidence Shortcut: A Reasoning Failure Mode of Masked Diffusion Models
von: Kim, Dueun, et al.
Veröffentlicht: (2026)

Dataforge: Agentic Platform for Autonomous Data Engineering
von: Wang, Xinyuan, et al.
Veröffentlicht: (2025)

DR.GAP: Mitigating Bias in Large Language Models using Gender-Aware Prompting with Demonstration and Reasoning
von: Qiu, Hongye, et al.
Veröffentlicht: (2025)

Mitigating Hallucination in Large Language Models (LLMs): An Application-Oriented Survey on RAG, Reasoning, and Agentic Systems
von: Li, Yihan, et al.
Veröffentlicht: (2025)

Blind Spot Navigation in Large Language Model Reasoning with Thought Space Explorer
von: Zhang, Jinghan, et al.
Veröffentlicht: (2024)

Thinking Fast, Thinking Wrong: Intuitiveness Modulates LLM Counterfactual Reasoning in Policy Evaluation
von: He, Yanjie
Veröffentlicht: (2026)

Stop Before You Fail: Operational Capability Boundaries for Mitigating Unproductive Reasoning in Large Reasoning Models
von: Zhang, Qingjie, et al.
Veröffentlicht: (2025)

To Think or Not To Think, That is The Question for Large Reasoning Models in Theory of Mind Tasks
von: Gong, Nanxu, et al.
Veröffentlicht: (2026)

Context-Aware Counterfactual Data Augmentation for Gender Bias Mitigation in Language Models
von: Parihar, Shweta, et al.
Veröffentlicht: (2026)

Locating and Mitigating Gender Bias in Large Language Models
von: Cai, Yuchen, et al.
Veröffentlicht: (2024)

Robust Search with Uncertainty-Aware Value Models for Language Model Reasoning
von: Yu, Fei, et al.
Veröffentlicht: (2025)

Logical Reasoning in Large Language Models: A Survey
von: Liu, Hanmeng, et al.
Veröffentlicht: (2025)

InfiR2: A Comprehensive FP8 Training Recipe for Reasoning-Enhanced Language Models
von: Wang, Wenjun, et al.
Veröffentlicht: (2025)

Chimera: Diagnosing Shortcut Learning in Visual-Language Understanding
von: Chi, Ziheng, et al.
Veröffentlicht: (2025)

Towards Concise and Adaptive Thinking in Large Reasoning Models: A Survey
von: Zhu, Jason, et al.
Veröffentlicht: (2025)

Shattering the Shortcut: A Topology-Regularized Benchmark for Multi-hop Medical Reasoning in LLMs
von: Zi, Xing, et al.
Veröffentlicht: (2026)

Confidence-Aware Routing for Large Language Model Reliability Enhancement: A Multi-Signal Approach to Pre-Generation Hallucination Mitigation
von: M, Nandakishor
Veröffentlicht: (2025)

Entropy-based Exploration Conduction for Multi-step Reasoning
von: Zhang, Jinghan, et al.
Veröffentlicht: (2025)

Token-Level Uncertainty-Aware Objective for Language Model Post-Training
von: Liu, Tingkai, et al.
Veröffentlicht: (2025)

Source-Aware Training Enables Knowledge Attribution in Language Models
von: Khalifa, Muhammad, et al.
Veröffentlicht: (2024)

The Dog the Cat Chased Stumped the Model: Measuring When Language Models Abandon Structure for Shortcuts
von: Madhusudan, Sangmitra, et al.
Veröffentlicht: (2025)

Learning Shortcuts: On the Misleading Promise of NLU in Language Models
von: Bihani, Geetanjali, et al.
Veröffentlicht: (2024)

TS-Reasoner: Aligning Time Series Foundation Models with LLM Reasoning
von: Yu, Fangxu, et al.
Veröffentlicht: (2025)

Learn More, Forget Less: A Gradient-Aware Data Selection Approach for LLM
von: Liu, Yibai, et al.
Veröffentlicht: (2025)

Knowing the Facts but Choosing the Shortcut: Understanding How Large Language Models Compare Entities
von: Lehmann, Hans Hergen, et al.
Veröffentlicht: (2025)

Coupled Variational Reinforcement Learning for Language Model General Reasoning
von: Wen, Xueru, et al.
Veröffentlicht: (2025)

BriLLM: Brain-inspired Large Language Model
von: Zhao, Hai, et al.
Veröffentlicht: (2025)

Crowdsourcing with Enhanced Data Quality Assurance: An Efficient Approach to Mitigate Resource Scarcity Challenges in Training Large Language Models for Healthcare
von: Barai, P., et al.
Veröffentlicht: (2024)

Mitigating Hallucinations in Large Language Models via Causal Reasoning
von: Li, Yuangang, et al.
Veröffentlicht: (2025)

Investigating Thinking Behaviours of Reasoning-Based Language Models for Social Bias Mitigation
von: Luo, Guoqing, et al.
Veröffentlicht: (2025)

Intent-Aware Self-Correction for Mitigating Social Biases in Large Language Models
von: Anantaprayoon, Panatchakorn, et al.
Veröffentlicht: (2025)