Cover Image

Soft Thinking: Unlocking the Reasoning Potential of LLMs in Continuous Concept Space

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Zhang, Zhen, He, Xuehai, Yan, Weixiang, Shen, Ao, Zhao, Chenyang, Wang, Shuohang, Shen, Yelong, Wang, Xin Eric
Format:	Preprint
Published:	2025
Subjects:	Computation and Language Artificial Intelligence
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

LoRC: Low-Rank Compression for LLMs KV Cache with a Progressive Compression Strategy
by: Zhang, Rongzhi, et al.
Published: (2024)

Reinforcement Learning for Reasoning in Large Language Models with One Training Example
by: Wang, Yiping, et al.
Published: (2025)

Mojito: Motion Trajectory and Intensity Control for Video Generation
by: He, Xuehai, et al.
Published: (2024)

Is Your World Simulator a Good Story Presenter? A Consecutive Events-Based Benchmark for Future Long Video Generation
by: Wang, Yiping, et al.
Published: (2024)

Zero Token-Driven Deep Thinking in LLMs: Unlocking the Full Potential of Existing Parameters via Cyclic Refinement
by: Li, Guanghao, et al.
Published: (2025)

Unlocking Recursive Thinking of LLMs: Alignment via Refinement
by: Zhang, Haoke, et al.
Published: (2025)

Efficient Reasoning with Hidden Thinking
by: Shen, Xuan, et al.
Published: (2025)

LLMs are Single-threaded Reasoners: Demystifying the Working Mechanism of Soft Thinking
by: Wu, Junhong, et al.
Published: (2025)

Cost-Effective Proxy Reward Model Construction with On-Policy and Active Learning
by: Chen, Yifang, et al.
Published: (2024)

CopT: Contrastive On-Policy Thinking with Continuous Spaces for General and Agentic Reasoning
by: Shi, Dachuan, et al.
Published: (2026)

KAG-Thinker: Interactive Thinking and Deep Reasoning in LLMs via Knowledge-Augmented Generation
by: Zhang, Dalong, et al.
Published: (2025)

CoDAR: Continuous Diffusion Language Models are More Powerful Than You Think
by: Shen, Junzhe, et al.
Published: (2026)

Latent Recurrent Transformer: Architecture Exploration, Training Strategies, and Scaling Behavior
by: Huang, Zeyi, et al.
Published: (2026)

Routing Mamba: Scaling State Space Models with Mixture-of-Experts Projection
by: Zhan, Zheng, et al.
Published: (2025)

Multi-LoRA Composition for Image Generation
by: Zhong, Ming, et al.
Published: (2024)

Temperature-Centric Investigation of Speculative Decoding with Knowledge Distillation
by: Ouyang, Siru, et al.
Published: (2024)

ThinkRouter: Efficient Reasoning via Routing Thinking between Latent and Discrete Spaces
by: Xu, Xin, et al.
Published: (2026)

StepFun-Formalizer: Unlocking the Autoformalization Potential of LLMs through Knowledge-Reasoning Fusion
by: Wu, Yutong, et al.
Published: (2025)

Read Quietly, Think Aloud: Decoupling Comprehension and Reasoning in LLMs
by: Wang, Yuanxin, et al.
Published: (2025)

ComCLIP: Training-Free Compositional Image and Text Matching
by: Jiang, Kenan, et al.
Published: (2022)

GRIT: Teaching MLLMs to Think with Images
by: Fan, Yue, et al.
Published: (2025)

Beyond Single Concept Vector: Modeling Concept Subspace in LLMs with Gaussian Distribution
by: Zhao, Haiyan, et al.
Published: (2024)

Draft-Thinking: Learning Efficient Reasoning in Long Chain-of-Thought LLMs
by: Cao, Jie, et al.
Published: (2026)

When to Continue Thinking: Adaptive Thinking Mode Switching for Efficient Reasoning
by: Zhang, Xiaoyun, et al.
Published: (2025)

Key-Point-Driven Data Synthesis with its Enhancement on Mathematical Reasoning
by: Huang, Yiming, et al.
Published: (2024)

MR-Ben: A Meta-Reasoning Benchmark for Evaluating System-2 Thinking in LLMs
by: Zeng, Zhongshen, et al.
Published: (2024)

MiMo: Unlocking the Reasoning Potential of Language Model -- From Pretraining to Posttraining
by: Xiaomi, LLM-Core, et al.
Published: (2025)

Unlocking LLM Creativity in Science through Analogical Reasoning
by: Shen, Andrew, et al.
Published: (2026)

Adapting LLM Agents with Universal Feedback in Communication
by: Wang, Kuan, et al.
Published: (2023)

Think Carefully and Check Again! Meta-Generation Unlocking LLMs for Low-Resource Cross-Lingual Summarization
by: Li, Zhecheng, et al.
Published: (2024)

ASPD: Unlocking Adaptive Serial-Parallel Decoding by Exploring Intrinsic Parallelism in LLMs
by: Chen, Keyu, et al.
Published: (2025)

Reverse Thinking Makes LLMs Stronger Reasoners
by: Chen, Justin Chih-Yao, et al.
Published: (2024)

JARVIS: A Neuro-Symbolic Commonsense Reasoning Framework for Conversational Embodied Agents
by: Zheng, Kaizhi, et al.
Published: (2022)

Q-Mirror: Unlocking the Multi-Modal Potential of Scientific Text-Only QA Pairs
by: Wang, Junying, et al.
Published: (2025)

ToRA: A Tool-Integrated Reasoning Agent for Mathematical Problem Solving
by: Gou, Zhibin, et al.
Published: (2023)

Scaling Up RL: Unlocking Diverse Reasoning in LLMs via Prolonged Training
by: Liu, Mingjie, et al.
Published: (2025)

InftyThink: Breaking the Length Limits of Long-Context Reasoning in Large Language Models
by: Yan, Yuchen, et al.
Published: (2025)

Enhancing Safety of Large Language Models via Embedding Space Separation
by: Zhao, Xu, et al.
Published: (2026)

Efficient Reasoning with Balanced Thinking
by: Li, Yulin, et al.
Published: (2026)

Improving Latent Reasoning in LLMs via Soft Concept Mixing
by: Wang, Kang, et al.
Published: (2025)