English
Related papers

Related papers: Rethinking Reflection in Pre-Training

200 papers

Why do thinking language models like DeepSeek R1 outperform their base counterparts? Despite consistent performance gains, it remains unclear to what extent thinking models learn entirely new reasoning capabilities or repurpose pre-existing…

Artificial Intelligence · Computer Science 2025-10-23 Constantin Venhoff , Iván Arcuschin , Philip Torr , Arthur Conmy , Neel Nanda

Large language models have recently demonstrated remarkable abilities to self-correct their responses through iterative refinement, often referred to as self-consistency or self-reflection. However, the dynamics of this self-correction…

Computation and Language · Computer Science 2025-11-13 Hossein A. Rahmani , Satyapriya Krishna , Xi Wang , Mohammadmehdi Naghiaei , Emine Yilmaz

Self-improvement through post-training methods such as iterative preference learning has been acclaimed for enhancing the problem-solving capabilities (e.g., mathematical reasoning) of Large Language Models (LLMs) without human…

Computation and Language · Computer Science 2024-07-09 Ting Wu , Xuefeng Li , Pengfei Liu

In this work, we investigate how explicitly modeling problem's difficulty prior information shapes the effectiveness of reinforcement learning based fine-tuning for multimodal reasoning. Our exploration mainly comprises of following three…

Computer Vision and Pattern Recognition · Computer Science 2025-12-16 Mingrui Chen , Haogeng Liu , Hao Liang , Huaibo Huang , Wentao Zhang , Ran He

Large Language Models (LLMs) often produce plausible but poorly-calibrated answers, limiting their reliability on reasoning-intensive tasks. We present Reinforcement Learning from Self-Feedback (RLSF), a post-training stage that uses the…

Computation and Language · Computer Science 2025-07-30 Carel van Niekerk , Renato Vukovic , Benjamin Matthias Ruppik , Hsien-chin Lin , Milica Gašić

Scaling model size and training data has led to great advances in the performance of Large Language Models (LLMs). However, the diminishing returns of this approach necessitate alternative methods to improve model capabilities, particularly…

Machine Learning · Computer Science 2025-11-05 Daman Arora , Andrea Zanette

Large Language Models (LLMs) solely trained on next-token prediction learn to solve a wide range of problems involving mathematical reasoning. But how does this ability evolve during training? We show the first analysis of how mathematical…

Artificial Intelligence · Computer Science 2025-12-15 Shubhra Mishra , Gabriel Poesia , Noah D. Goodman

Chain-of-thought (CoT) reasoning has enabled large language models (LLMs) to utilize additional computation through intermediate tokens to solve complex tasks. However, we posit that typical reasoning traces contain many redundant tokens,…

Computation and Language · Computer Science 2025-06-11 Tergel Munkhbat , Namgyu Ho , Seo Hyun Kim , Yongjin Yang , Yujin Kim , Se-Young Yun

Models trained on a new task typically degrade on prior tasks, a phenomenon known as forgetting. Traditionally, mitigating forgetting has required replaying stored exemplars from prior tasks, which is often impractical. By contrast,…

Machine Learning · Computer Science 2026-05-26 Martin Marek , Dongkyu Cho , Shikai Qiu , Rumi Chunara , Pavel Izmailov , Andrew Gordon Wilson

While deep reasoning with long chain-of-thought has dramatically improved large language models in verifiable domains like mathematics, its effectiveness for open-ended tasks such as writing remains unexplored. In this paper, we conduct a…

Computation and Language · Computer Science 2026-04-06 Wanlong Liu , Bo Zhang , Chenliang Li , Shaopeng Lai , Yuning Wu , Xuanyu Lei , Ming Yan

Self-improvement is a mechanism in Large Language Model (LLM) pre-training, post-training and test-time inference. We explore a framework where the model verifies its own outputs, filters or reweights data based on this verification, and…

Computation and Language · Computer Science 2025-02-26 Yuda Song , Hanlin Zhang , Carson Eisenach , Sham Kakade , Dean Foster , Udaya Ghai

Recent studies on transformer-based language models show that they can answer questions by reasoning over knowledge provided as part of the context (i.e., in-context reasoning). However, since the available knowledge is often not filtered…

Computation and Language · Computer Science 2023-11-07 Zeming Chen , Gail Weiss , Eric Mitchell , Asli Celikyilmaz , Antoine Bosselut

Reasoning Large Language Models (RLLMs) have demonstrated impressive performance on complex tasks, largely due to the adoption of Long Chain-of-Thought (Long CoT) reasoning. However, they often exhibit overthinking -- performing unnecessary…

Computation and Language · Computer Science 2025-05-30 Keqin Peng , Liang Ding , Yuanxin Ouyang , Meng Fang , Dacheng Tao

Reinforcement learning is now widely adopted as the final stage of large language model training, especially for reasoning-style tasks such as maths problems. Typically, models attempt each question many times during a single training step…

Machine Learning · Computer Science 2026-05-01 Thomas Foster , Anya Sims , Johannes Forkel , Mattie Fellows , Jakob Foerster

Recent studies suggest that self-reflective prompting can significantly enhance the reasoning capabilities of Large Language Models (LLMs). However, the use of external feedback as a stop criterion raises doubts about the true extent of…

Computation and Language · Computer Science 2024-04-16 Yanhong Li , Chenghao Yang , Allyson Ettinger

Recent work on language model self-improvement shows that models can refine their own reasoning through reflection, verification, debate, or self-generated rewards. However, most existing approaches rely on external critics, learned reward…

Artificial Intelligence · Computer Science 2026-01-06 Mandar Parab

Multimodal Large Language Models (MLLMs) equipped with step-by-step thinking capabilities have demonstrated remarkable performance on complex reasoning problems. However, this thinking process is redundant for simple problems solvable…

Computer Vision and Pattern Recognition · Computer Science 2025-09-03 Qi Yang , Bolin Ni , Shiming Xiang , Han Hu , Houwen Peng , Jie Jiang

Large language models excel on static benchmarks, but their ability as self-learning agents in dynamic environments remains unclear. We evaluate three prompting strategies: self-reflection, heuristic mutation, and planning across dynamic…

Artificial Intelligence · Computer Science 2025-08-12 Annie Wong , Thomas Bäck , Aske Plaat , Niki van Stein , Anna V. Kononova

Large Reasoning Models possess remarkable capabilities for self-correction in general domain; however, they frequently struggle to recover from unsafe reasoning trajectories under adversarial attacks. Existing alignment methods attempt to…

Artificial Intelligence · Computer Science 2026-05-12 Dongcheng Zhang , Yi Zhang , Yuxin Chen , An Zhang , Xiang Wang , Chaochao Lu

Through reinforcement learning (RL) with outcome correctness rewards, large reasoning models (LRMs) with scaled inference computation have demonstrated substantial success on complex reasoning tasks. However, the one-sided reward, focused…

Computation and Language · Computer Science 2025-11-21 Jiashu Yao , Heyan Huang , Shuang Zeng , Chuwei Luo , WangJie You , Jie Tang , Qingsong Liu , Yuhang Guo , Yangyang Kang