中文
相关论文

相关论文: Quagmires in SFT-RL Post-Training: When High SFT S…

200 篇论文

Recent advances in vision-language models (VLMs) reasoning have been largely attributed to the rise of reinforcement Learning (RL), which has shifted the community's focus away from the supervised fine-tuning (SFT) paradigm. Many studies…

Pre-trained Large Language Model (LLM) exhibits broad capabilities, yet, for specific tasks or domains their attainment of higher accuracy and more reliable reasoning generally depends on post-training through Supervised Fine-Tuning (SFT)…

人工智能 · 计算机科学 2026-03-17 Haitao Jiang , Wenbo Zhang , Jiarui Yao , Hengrui Cai , Sheng Wang , Rui Song

Large Language Models (LLMs) are pretrained on massive datasets and later instruction-tuned via supervised fine-tuning (SFT) or reinforcement learning (RL). Best practices emphasize large, diverse pretraining data, whereas post-training…

机器学习 · 计算机科学 2026-03-03 Adel Javanmard , Baharan Mirzasoleiman , Vahab Mirrokni

Reinforcement learning (RL)-based post-training often improves the reasoning performance of large language models (LLMs) beyond the training domain, while supervised fine-tuning (SFT) frequently leads to general capabilities forgetting.…

计算与语言 · 计算机科学 2026-04-29 Dan Shi , Zhuowen Han , Simon Ostermann , Renren Jin , Josef van Genabith , Deyi Xiong

Large language models (LLMs) trained via pretraining and supervised fine-tuning (SFT) can still produce harmful and misaligned outputs, or struggle in domains like math and coding. Reinforcement learning (RL)-based post-training methods,…

Supervised Fine-Tuning (SFT) followed by Reinforcement Learning (RL) is a standard post-training recipe for improving Large Language Models (LLM) reasoning, but why it works remains unclear. We revisit the common claim that ``SFT memorizes,…

Post-training of large language models routinely interleaves supervised fine-tuning (SFT) with reinforcement learning (RL). These two methods have different objectives: SFT minimizes the cross-entropy loss between model outputs and expert…

机器学习 · 计算机科学 2026-05-07 Xueyan Niu , Bo Bai , Wei Han , Weixi Zhang

Large vision-language models (VLMs) increasingly adopt post-training techniques such as long chain-of-thought (CoT) supervised fine-tuning (SFT) and reinforcement learning (RL) to elicit sophisticated reasoning. While these methods exhibit…

计算与语言 · 计算机科学 2025-07-11 Jierun Chen , Tiezheng Yu , Haoli Bai , Lewei Yao , Jiannan Wu , Kaican Li , Fei Mi , Chaofan Tao , Lei Zhu , Manyi Zhang , Xiaohui Li , Lu Hou , Lifeng Shang , Qun Liu

Training large language models (LLMs) for reasoning via maths and code datasets has become a major new focus in LLM post-training. Two particularly popular approaches are reinforcement learning (RL) and supervised fine-tuning (SFT), but…

机器学习 · 计算机科学 2025-07-28 Neel Rajani , Aryo Pradipta Gema , Seraphina Goldfarb-Tarrant , Ivan Titov

Training large language models (LLMs) from scratch is increasingly impractical, making post-training methods such as supervised fine-tuning (SFT) and reinforcement-learning fine-tuning (RL-FT, e.g., PPO) central to modern practice. Using an…

机器学习 · 计算机科学 2025-08-25 Hangzhan Jin , Sicheng Lv , Sifan Wu , Mohammad Hamdaqa

Large Language Models (LLMs) show strong reasoning abilities, often amplified by Chain-of-Thought (CoT) prompting and reinforcement learning (RL). Although RL algorithms can substantially improve reasoning, they struggle to expand reasoning…

计算与语言 · 计算机科学 2025-10-07 Xiangchi Yuan , Xiang Chen , Tong Yu , Dachuan Shi , Can Jin , Wenke Lee , Saayan Mitra

Supervised fine-tuning (SFT) and reinforcement learning (RL) are widely used post-training techniques for foundation models. However, their roles in enhancing model generalization capabilities remain unclear. This paper studies the…

人工智能 · 计算机科学 2025-05-27 Tianzhe Chu , Yuexiang Zhai , Jihan Yang , Shengbang Tong , Saining Xie , Dale Schuurmans , Quoc V. Le , Sergey Levine , Yi Ma

This work revisits the dominant supervised fine-tuning (SFT) then reinforcement learning (RL) paradigm for training Large Vision-Language Models (LVLMs), and reveals a key finding: SFT can significantly undermine subsequent RL by inducing…

计算与语言 · 计算机科学 2025-04-17 Hardy Chen , Haoqin Tu , Fali Wang , Hui Liu , Xianfeng Tang , Xinya Du , Yuyin Zhou , Cihang Xie

Improving Multi-modal Large Language Models (MLLMs) in the post-training stage typically relies on supervised fine-tuning (SFT) or reinforcement learning (RL), which require expensive and manually annotated multi-modal data--an ultimately…

计算与语言 · 计算机科学 2025-10-28 Lai Wei , Yuting Li , Chen Wang , Yue Wang , Linghe Kong , Weiran Huang , Lichao Sun

After the pretraining stage of LLMs, techniques such as SFT, RLHF, RLVR, and RFT are applied to enhance instruction-following ability, mitigate undesired responses, improve reasoning capability and enable efficient domain adaptation with…

计算与语言 · 计算机科学 2025-10-17 Zhichao Wang , Andy Wong , Ruslan Belkin

Enhancing the complex reasoning capabilities of Large Language Models (LLMs) attracts widespread attention. While reinforcement learning (RL) has shown superior performance for improving complex reasoning, its impact on cross-lingual…

计算与语言 · 计算机科学 2025-09-30 Shulin Huang , Yiran Ding , Junshu Pan , Yue Zhang

Fine-tuning Large Language Models (LLMs) on specific datasets is a common practice to improve performance on target tasks. However, this performance gain often leads to overfitting, where the model becomes too specialized in either the task…

计算与语言 · 计算机科学 2024-09-10 Sonam Gupta , Yatin Nandwani , Asaf Yehudai , Mayank Mishra , Gaurav Pandey , Dinesh Raghu , Sachindra Joshi

Automatic code generation has been a longstanding research topic. With the advancement of general-purpose large language models (LLMs), the ability to code stands out as one important measure to the model's reasoning performance. Usually, a…

软件工程 · 计算机科学 2024-12-18 Jie Chen , Xintian Han , Yu Ma , Xun Zhou , Liang Xiang

Recent advances in large language model (LLM) reasoning have shown that sophisticated behaviors such as planning and self-reflection can emerge through reinforcement learning (RL). However, despite these successes, RL in its current form…

Mathematical reasoning is a challenging task for large language models (LLMs), while the scaling relationship of it with respect to LLM capacity is under-explored. In this paper, we investigate how the pre-training loss, supervised data…

计算与语言 · 计算机科学 2023-09-14 Zheng Yuan , Hongyi Yuan , Chengpeng Li , Guanting Dong , Keming Lu , Chuanqi Tan , Chang Zhou , Jingren Zhou
‹ 上一页 1 2 3 10 下一页 ›