Cover Image

Latent Reward Steering: An Adaptive Inference-Time Framework that Implicitly Promotes Cognitive Behaviors in Reasoning LLMs

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Li, Jiakang, Zhu, Guanyu, Jin, Can, Huang, Chenxi, Yu, Dexu, Chen, Ronghao, Zhou, Yang, Peng, Hongwu, Lan, Xuanqi, Metaxas, Dimitris N., Li, Youhua
Format:	Preprint
Published:	2026
Subjects:	Artificial Intelligence
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Weak Critics Make Strong Learners: On-Policy Critique Distillation for Scalable Oversight
by: Jin, Can, et al.
Published: (2026)

Two Heads are Better Than One: Test-time Scaling of Multi-agent Collaborative Reasoning
by: Jin, Can, et al.
Published: (2025)

Test-Time Spectrum-Aware Latent Steering for Zero-Shot Generalization in Vision-Language Models
by: Dafnis, Konstantinos M., et al.
Published: (2025)

Learning from Teaching Regularization: Generalizable Correlations Should be Easy to Imitate
by: Jin, Can, et al.
Published: (2024)

GSS: Gated Subspace Steering for Selective Memorization Mitigation in LLMs
by: Zhang, Xuanqi, et al.
Published: (2026)

Chain of Mindset: Reasoning with Adaptive Cognitive Modes
by: Jiang, Tianyi, et al.
Published: (2026)

Thinking in Latents: Adaptive Anchor Refinement for Implicit Reasoning in LLMs
by: Sheshanarayana, Disha, et al.
Published: (2026)

Your Reward Function for RL is Your Best PRM for Search: Unifying RL and Search-Based TTS
by: Jin, Can, et al.
Published: (2025)

Beyond Explicit Edges: Robust Reasoning over Noisy and Sparse Knowledge Graphs
by: Gao, Hang, et al.
Published: (2026)

On the Role of Language Representations in Auto-Bidding: Findings and Implications
by: Zhu, Guanyu, et al.
Published: (2026)

Steering Rectified Flow Models in the Vector Field for Controlled Image Generation
by: Patel, Maitreya, et al.
Published: (2024)

RISER: Orchestrating Latent Reasoning Skills for Adaptive Activation Steering
by: Ye, Wencheng, et al.
Published: (2026)

ATLAS: Adaptive Test-Time Latent Steering with External Verifiers for Enhancing LLMs Reasoning
by: Nguyen, Tuc, et al.
Published: (2026)

Beyond Interpretability: When, Why, and How Sparse Autoencoders Enable Label-Free Visual Steering
by: Chatzoudis, Gerasimos, et al.
Published: (2025)

Distribution-Aware Reward Estimation for Test-Time Reinforcement Learning
by: Du, Bodong, et al.
Published: (2026)

MPDiT: Multi-Patch Global-to-Local Transformer Architecture For Efficient Flow Matching and Diffusion Model
by: Dao, Quan, et al.
Published: (2026)

Latent Implicit Visual Reasoning
by: Li, Kelvin, et al.
Published: (2025)

Implicit In-context Learning
by: Li, Zhuowei, et al.
Published: (2024)

Improved Training Technique for Latent Consistency Models
by: Dao, Quan, et al.
Published: (2025)

Understanding and Steering the Cognitive Behaviors of Reasoning Models at Test-Time
by: Zhang, Zhenyu, et al.
Published: (2025)

Reasoning over Precedents Alongside Statutes: Case-Augmented Deliberative Alignment for LLM Safety
by: Jin, Can, et al.
Published: (2026)

CogDual: Enhancing Dual Cognition of LLMs via Reinforcement Learning with Implicit Rule-Based Rewards
by: Liu, Cheng, et al.
Published: (2025)

Reinforcement Learning with Verifiable Rewards Implicitly Incentivizes Correct Reasoning in Base LLMs
by: Wen, Xumeng, et al.
Published: (2025)

RankFlow: A Multi-Role Collaborative Reranking Workflow Utilizing Large Language Models
by: Jin, Can, et al.
Published: (2025)

DTop-p MoE: Sparsity-Controlled Dynamic Top-p MoE for Foundation Model Pre-training
by: Jin, Can, et al.
Published: (2025)

APEER: Automatic Prompt Engineering Enhances Large Language Model Reranking
by: Jin, Can, et al.
Published: (2024)

Mitigating Cognitive Inertia in Large Reasoning Models via Latent Spike Steering
by: Lee, Seojin, et al.
Published: (2026)

Token-Controlled Re-ranking for Sequential Recommendation via LLMs
by: Dai, Wenxi, et al.
Published: (2025)

SteerConf: Steering LLMs for Confidence Elicitation
by: Zhou, Ziang, et al.
Published: (2025)

Fractional Reasoning via Latent Steering Vectors Improves Inference Time Compute
by: Liu, Sheng, et al.
Published: (2025)

FairSteer: Inference Time Debiasing for LLMs with Dynamic Activation Steering
by: Li, Yichen, et al.
Published: (2025)

DARE: Difficulty-Adaptive Reinforcement Learning with Co-Evolved Difficulty Estimation
by: Zhou, Yang, et al.
Published: (2026)

iCLP: Large Language Model Reasoning with Implicit Cognition Latent Planning
by: Chen, Sijia, et al.
Published: (2025)

Stabilizing Knowledge, Promoting Reasoning: Dual-Token Constraints for RLVR
by: Wang, Jiakang, et al.
Published: (2025)

LED: LLM Enhanced Open-Vocabulary Object Detection without Human Curated Data Generation
by: Zhou, Yang, et al.
Published: (2025)

Exploring the Personality Traits of LLMs through Latent Features Steering
by: Yang, Shu, et al.
Published: (2024)

Cognitive Aesthetic Evaluation Model - Implementation Code
by: Chenxi, Jin
Published: (2026)

Improving Visual Reasoning with Iterative Evidence Refinement
by: Shi, Zeru, et al.
Published: (2026)

Score-Guided Diffusion for 3D Human Recovery
by: Stathopoulos, Anastasis, et al.
Published: (2024)

Adaptive Fuzzy‐Based Event‐Triggered Consensus of Switched Nonlinear Multiagent Systems With Communication Faults and State‐Dependent Switchings
by: Ronghao Zhang, et al.
Published: (2025)