Beyond ’Aha!’: Toward Systematic Meta-Abilities Alignment in Large Reasoning Models (2026.findings-acl)
Copied to clipboard
| Challenge: | Prior work has shown that outcome-based reinforcement learning (RL) can incidentally elicit advanced reasoning behaviors such as self-correction, backtracking, and verification. |
| Approach: | They explicitly align large reasoning models with three meta-abilities: deduction, induction, and abduction, using automatically generated, self-verifiable tasks. |
| Outcome: | The proposed model aligns models with deduction, induction, and abduction meta-abilities using automatically generated, self-verifiable tasks. |
Similar Papers
From "Aha Moments" to Controllable Thinking: Toward Meta-Cognitive Reasoning in LRMs via Decoupled Reasoning and Control (2026.acl-long)
Copied to clipboard
| Challenge: | Large Reasoning Models exhibit step-by-step reasoning, reflection, and backtracking, but these behaviors are often unregulated, leading to overthinking. |
| Approach: | They propose a meta-cognitive reasoning framework that decouples reasoning from control to enable independent optimization of control strategies. |
| Outcome: | Experiments show that the proposed model improves efficiency and accuracy across reasoning benchmarks. |
Towards a Mechanistic Understanding of Large Reasoning Models: A Survey of Training, Inference, and Failures (2026.acl-long)
Copied to clipboard
Yi Hu, Jiaqi Gu, Ruxin Wang, Zijun Yao, Hao Peng, Xiaobao Wu, Jianhui Chen, Muhan Zhang, Liangming Pan
| Challenge: | Recent research has shown that reinforcement learning can elicit intriguing emergent reasoning behaviors. |
| Approach: | They propose a comprehensive survey of the mechanistic understanding of large reasoning models . they organize findings into three core dimensions: 1) training dynamics, 2) reasoning mechanisms, and 3) unintended behaviors. |
| Outcome: | This paper synthesizes the mechanistic understanding of large reasoning models into three dimensions . authors outline a roadmap for future studies including improved interpretability and methodologies . |
PARIF: Pushing the Pareto Frontier of Instruction Following and Reasoning with Curriculum Reinforcement Learning (2026.acl-long)
Copied to clipboard
| Challenge: | Existing alignment methods struggle to balance general reasoning with instruction-following (IF) this is hindered by dependency on teacher models, reward hacking, and reasoning-answer inconsistencies. |
| Approach: | They propose a two-stage curriculum learning framework based on Reinforcement Learning from Verifiable Rewards to enhance both IF and general reasoning capabilities. |
| Outcome: | The proposed framework outperforms leading models on six representative IF tasks while achieving a 21.25% relative average improvement over the original model. |
Reasoning Structure Matters for Safety Alignment of Reasoning Models (2026.acl-long)
Copied to clipboard
| Challenge: | Large reasoning models (LRMs) achieve strong performance on complex reasoning tasks but often generate harmful responses to malicious user queries. |
| Approach: | They propose a method that alters the reasoning structure of large reasoning models to achieve effective safety alignment by supervised fine-tuning. |
| Outcome: | The proposed method is practical and generalizable, requiring no complex training or reward design. |
AlgBench: To What Extent Do Large Reasoning Models Understand Algorithms? (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing benchmarks for algorithmic reasoning fail to answer a critical question: do LRMs master algorithmic thinking? Empirical evaluations on leading LRM models reveal substantial performance heterogeneity, while models perform well on non-optimized tasks, accuracy drops sharply to around 49% on globally optimized algorithms. |
| Approach: | They propose an algorithm-centric benchmark that evaluates large reasoning models under an algorithmic paradigm. |
| Outcome: | Empirical evaluations on leading LRMs reveal substantial performance heterogeneity . models perform well on non-optimized tasks, accuracy drops sharply to around 49% . |
RiT: Rubrics-in-Thinking Reinforcement Learning for Improved Reasoning in Large Language Models (2026.findings-acl)
Copied to clipboard
| Challenge: | Large Reasoning Models benefit from generating intermediate reasoning steps alongside final answers. |
| Approach: | They propose a framework to introduce thinking-rubric supervision into intermediate reasoning. |
| Outcome: | The proposed framework outperforms outcome-only RL baselines on reasoning-intensive and open-ended tasks. |
PAM: Enhancing General Alignment of Large Reasoning Models through Priority-Aware Metacognition (2026.acl-long)
Copied to clipboard
| Challenge: | Existing studies indicate that System-2 thinking alone does not transfer to the general alignment domain. |
| Approach: | They propose to use priority-aware metacognition to help LRMs understand human preferences and monitor and regulate their thinking process. |
| Outcome: | The proposed model improves general alignment performance by 10 points on helpfulness and harmless benchmarks. |
MR-ALIGN: Meta-Reasoning Informed Factuality Alignment for Large Reasoning Models (2026.findings-acl)
Copied to clipboard
Xinming Wang, Jian Xu, Bin Yu, Sheng Lian, yi Chen, Boran Wang, Yingjian Zhu, Hongzhu Yi, Hong-Ming Yang, Han Hu, Cheng-Lin Liu, Xu-Yao Zhang
| Challenge: | Large reasoning models (LRMs) show strong capabilities in complex reasoning, yet their marginal gains on evidence-dependent factual questions are limited. |
| Approach: | They propose a Meta-Reasoning informed alignment framework that quantifies state-transition probabilities along the model’s thinking process and constructs a transition-aware implicit reward that reinforces beneficial reasoning patterns while suppressing defective ones at the atomic thinking segments. |
| Outcome: | Empirical evaluations of four factual QA datasets and one long-form factuality benchmark show that MR-ALIGN consistently improves accuracy and truthfulness while reducing misleading reasoning. |
Inverse Reinforcement Learning Meets Large Language Model Alignment (2025.acl-tutorials)
Copied to clipboard
| Challenge: | This tutorial will provide a comprehensive review of recent advances in LLM alignment . it will highlight the necessity of constructing neural reward models from human data . |
| Approach: | This tutorial will provide a comprehensive review of recent advances in LLM alignment through the lens of inverse reinforcement learning. |
| Outcome: | This tutorial will provide a comprehensive review of recent advances in LLM alignment through the lens of inverse reinforcement learning (IRL). |
The Illusion of Insight in Reasoning Models (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing evidence suggests that reasoning models exhibit "Aha!" moments . however, it remains unclear whether such intrinsic shifts in reasoning strategy improve performance . |
| Approach: | They analyze reasoning traces, training checks, three reasoning domains to detect mid-reasoning shifts . they find that reasoning shifts are rare, do not become more frequent with training . |
| Outcome: | The proposed method detects mid-reasoning shifts and instrument training runs. |