Image de couverture de livre

Beyond Negative Rollouts: Positive-Only Policy Optimization with Implicit Negative Gradients

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteurs principaux:	Xu, Mingwei, Fang, Hao
Format:	Preprint
Publié:	2026
Sujets:	Computation and Language
Accès en ligne:	Acceder al recurso
Tags:	Ajouter un tag Pas de tags, Soyez le premier à ajouter un tag!

Documents similaires

On the Effect of Negative Gradient in Group Relative Deep Reinforcement Optimization
par: Deng, Wenlong, et autres
Publié: (2025)

Negating Negatives: Alignment with Human Negative Samples via Distributional Dispreference Optimization
par: Duan, Shitong, et autres
Publié: (2024)

iACOS: Advancing Implicit Sentiment Extraction with Informative and Adaptive Negative Examples
par: Xu, Xiancai, et autres
Publié: (2023)

One-Token Rollout: Guiding Supervised Fine-Tuning of LLMs with Policy Gradient
par: Ming, Rui, et autres
Publié: (2025)

Are Full Rollouts Necessary for On-Policy Distillation?
par: Zhang, Yaocheng, et autres
Publié: (2026)

Negative to Positive Co-learning with Aggressive Modality Dropout
par: Magal, Nicholas, et autres
Publié: (2025)

Negative Before Positive: Asymmetric Valence Processing in Large Language Models
par: Venkatesh, Sohan
Publié: (2026)

Not All Rollouts are Useful: Down-Sampling Rollouts in LLM Reinforcement Learning
par: Xu, Yixuan Even, et autres
Publié: (2025)

NegativePrompt: Leveraging Psychology for Large Language Models Enhancement via Negative Emotional Stimuli
par: Wang, Xu, et autres
Publié: (2024)

Commonsense Knowledge with Negation: A Resource to Enhance Negation Understanding
par: Wang, Zijie, et autres
Publié: (2026)

Carrot and Stick: Inducing Self-Motivation with Positive & Negative Feedback
par: Sohn, Jimin, et autres
Publié: (2024)

Enhancing AI Assisted Writing with One-Shot Implicit Negative Feedback
par: Towle, Benjamin, et autres
Publié: (2024)

Unintended Negative Impacts of Promotional Language in Patent Evaluation
par: Zhao, Bingkun, et autres
Publié: (2026)

Label-aware Hard Negative Sampling Strategies with Momentum Contrastive Learning for Implicit Hate Speech Detection
par: Kim, Jaehoon, et autres
Publié: (2024)

DEO: Training-Free Direct Embedding Optimization for Negation-Aware Retrieval
par: Lee, Taegyeong, et autres
Publié: (2026)

Negation-Induced Forgetting in LLMs
par: Capuano, Francesca, et autres
Publié: (2025)

SPEC-RL: Accelerating On-Policy Reinforcement Learning with Speculative Rollouts
par: Liu, Bingshuai, et autres
Publié: (2025)

Beyond Token-Level Policy Gradients for Complex Reasoning with Large Language Models
par: Xu, Mufan, et autres
Publié: (2026)

How Language Models Process Negation
par: Zhou, Zhejian, et autres
Publié: (2026)

Learning Robust Negation Text Representations
par: Truong, Thinh Hung, et autres
Publié: (2025)

The Self-Contained Negation Test Set
par: Kletz, David, et autres
Publié: (2024)

Negative Matters: Multi-Granularity Hard-Negative Synthesis and Anchor-Token-Aware Pooling for Enhanced Text Embeddings
par: Pan, Tengyu, et autres
Publié: (2025)

Logical Negation Augmenting and Debiasing for Prompt-based Methods
par: Li, Yitian, et autres
Publié: (2024)

Skip-Connected Policy Optimization for Implicit Advantage
par: Teng, Fengwei, et autres
Publié: (2026)

Correcting Negative Bias in Large Language Models through Negative Attention Score Alignment
par: Yu, Sangwon, et autres
Publié: (2024)

Optimizing Legal Document Retrieval in Vietnamese with Semi-Hard Negative Mining
par: Le, Van-Hoang, et autres
Publié: (2025)

Importance Sampling for Multi-Negative Multimodal Direct Preference Optimization
par: Li, Xintong, et autres
Publié: (2025)

Paraphrasing in Affirmative Terms Improves Negation Understanding
par: Rezaei, MohammadHossein, et autres
Publié: (2024)

Unsupervised hard Negative Augmentation for contrastive learning
par: Shu, Yuxuan, et autres
Publié: (2024)

Greedy Grammar Induction with Indirect Negative Evidence
par: Potashnik, Joseph
Publié: (2023)

Making Language Models Robust Against Negation
par: Rezaei, MohammadHossein, et autres
Publié: (2025)

DreamArtist++: Controllable One-Shot Text-to-Image Generation via Positive-Negative Adapter
par: Dong, Ziyi, et autres
Publié: (2022)

Simplicity Prevails: Rethinking Negative Preference Optimization for LLM Unlearning
par: Fan, Chongyu, et autres
Publié: (2024)

Negative Preference Optimization: From Catastrophic Collapse to Effective Unlearning
par: Zhang, Ruiqi, et autres
Publié: (2024)

OptPO: Optimal Rollout Allocation for Test-time Policy Optimization
par: Wang, Youkang, et autres
Publié: (2025)

Negation Triplet Extraction with Syntactic Dependency and Semantic Consistency
par: Shi, Yuchen, et autres
Publié: (2024)

ROAST: Rollout-based On-distribution Activation Steering Technique
par: Su, Xuanbo, et autres
Publié: (2026)

SCENE: Self-Labeled Counterfactuals for Extrapolating to Negative Examples
par: Fu, Deqing, et autres
Publié: (2023)

Rethinking Negative Instances for Generative Named Entity Recognition
par: Ding, Yuyang, et autres
Publié: (2024)

Investigating and Addressing Hallucinations of LLMs in Tasks Involving Negation
par: Varshney, Neeraj, et autres
Publié: (2024)