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

arXiv: https://arxiv.org/abs/2603.20899 · HTML: https://mlsanity.com/paper/2603.20899/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Placed in Test-time compute and efficient thinking, one of 17 directions in LLM reasoning.
- Builds on 14 of its 20 references that are in the map.
- First posted on arXiv 21 Mar 2026.

## Placed in

- LLM reasoning / [Test-time compute and efficient thinking](https://mlsanity.com/lr/test-time-compute-and-efficient-thinking/)

## Builds on (most cited first, 14 of 14)

- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [Chain-of-Thought Prompting Elicits Reasoning in Large Language Models](https://mlsanity.com/paper/2201.11903/) (2022)
- [Self-Consistency Improves Chain of Thought Reasoning in Language Models](https://mlsanity.com/paper/2203.11171/) (2022)
- [Large Language Models are Zero-Shot Reasoners](https://mlsanity.com/paper/2205.11916/) (2022)
- [Trust Region Policy Optimization](https://mlsanity.com/paper/1502.05477/) (2015)
- [Invariant Risk Minimization](https://mlsanity.com/paper/1907.02893/) (2019)
- [Understanding Black-box Predictions via Influence Functions](https://mlsanity.com/paper/1703.04730/) (2017)
- [Gradient Surgery for Multi-Task Learning](https://mlsanity.com/paper/2001.06782/) (2020)
- [Distributionally Robust Neural Networks for Group Shifts: On the Importance of Regularization for Worst-Case Generalization](https://mlsanity.com/paper/1911.08731/) (2019)
- [Out-of-Distribution Generalization via Risk Extrapolation (REx)](https://mlsanity.com/paper/2003.00688/) (2020)
- [Just Train Twice: Improving Group Robustness without Training Group Information](https://mlsanity.com/paper/2107.09044/) (2021)
- [Large Language Models Are Reasoning Teachers](https://mlsanity.com/paper/2212.10071/) (2022)
- [Learning from Failure: Training Debiased Classifier from Biased Classifier](https://mlsanity.com/paper/2007.02561/) (2020)
- [A Survey of Reasoning with Foundation Models](https://mlsanity.com/paper/2312.11562/) (2023)

## Built on by (most cited first, 2 of 2)

- [Rethinking Data Augmentation under Covariate Shift: Invariant-Guided Diffusion and Prototype Reweighting](https://mlsanity.com/paper/2610.00873/) (2026)
- [Which and When to Admit: Gradient Admission for Data-Centric Small Language Model Finetuning](https://mlsanity.com/paper/2610.07553/) (2026)

## Abstract (arXiv)

Large language models exhibit strong reasoning capabilities, yet often rely on shortcuts such as surface pattern matching and answer memorization rather than genuine logical inference. We propose Shortcut-Aware Reasoning Training (SART), a gradient-aware framework that detects and mitigates shortcut-promoting samples via ShortcutScore and gradient surgery. Our method identifies shortcut signals through gradient misalignment with validation objectives and answer-token concentration, and modifies training dynamics accordingly. Experiments on controlled reasoning benchmarks show that SART achieves +16.5% accuracy and +40.2% robustness over the strongest baseline, significantly improving generalization under distribution shifts. Code is available at: https://github.com/fuyanjie/short-cut-aware-data-centric-reasoning.