SARL rewards reasoning topology to improve label-free RL, outperforming baselines with gains up to 44.7% on math and 34.6% on open-ended tasks while maintaining more stable training.
hub Mixed citations
Direct preference optimization: Your language model is secretly a reward model.Advances in neural information processing systems, 36:53728–53741
Mixed citation behavior. Most common role is background (67%).
18
Pith papers citing it
Background
67% of classified citations
hub tools
citation-role summary
background 4
method 2
citation-polarity summary
representative citing papers
RaPO reduces catastrophic forgetting in visual continual learning by shaping rewards around policy drift and stabilizing advantages with cross-task exponential moving averages during reinforcement fine-tuning of multimodal models.
BadDLM implants effective backdoors in diffusion language models across concept, attribute, alignment, and payload targets by exploiting denoising dynamics while preserving clean performance.
MSDDA derives a closed-form optimal reverse denoising distribution for multi-objective diffusion alignment that is exactly equivalent to step-level RL fine-tuning with no approximation error.
TTT-Discover applies test-time RL to set new state-of-the-art results on math inequalities, GPU kernels, algorithm contests, and single-cell denoising using an open model and public code.
Priority ranking offers a low-cost direct evaluation for harness optimizers that correlates with their real multi-step optimization performance, supported by the Shor dataset of 182 scenarios.
SURF derives weight sampling rules from the arc-length CDF of the scalarization path to uniformly traverse the Pareto front in multi-objective optimization.
LatentUMM proposes dual latent alignment at modality and capacity levels plus latent dynamics stabilization to reduce semantic drift and improve consistency in unified multimodal models.
Tail-extrapolated estimators approximate best-of-N policy gradients from limited training rollouts by leveraging upper-tail reward statistics under structural assumptions.
PIA achieves lower attack success rates on persona-based jailbreaks via self-play co-evolution of attacks (PLE) and defenses (PICL) that structurally decouples safety from persona context using unilateral KL-divergence.
A 17B-parameter sparse MoE diffusion transformer activates 2B parameters per pass and reaches competitive quality on image generation benchmarks without post-training.
Kling-Omni is a unified multimodal generative system that produces cinematic videos from diverse inputs by integrating generation, editing, and intelligent reasoning in a single end-to-end model.
Automated LLM-based prompt engineering for text-to-image edge-case synthesis improves object detection robustness on the FishEye8K benchmark over naive augmentation and manual prompts.
Negative-capable ridge regression uses controlled negative regularization as anti-shrinkage to increase effective complexity along weak eigendirections and mitigate underfitting in small-data regression.
PhyWorld improves temporal consistency and physical plausibility in video world models via flow matching fine-tuning followed by DPO on physics preference pairs, with reported gains on VBench and a custom physical-faithfulness benchmark.
The paper introduces the Proxy Compression Hypothesis as a unifying framework explaining reward hacking in RLHF as an emergent result of compressing high-dimensional human objectives into proxy reward signals under optimization pressure.
Z-Image is an efficient 6B-parameter foundation model for image generation that rivals larger commercial systems in photorealism and bilingual text rendering through a new single-stream diffusion transformer and streamlined training.
Position paper proposes replacing fragmented narrow AI models with LLMs as the cognitive orchestrator in the RAN Intelligent Controller for Level 5 autonomous 6G networks.
citing papers explorer
-
SARL: Label-Free Reinforcement Learning by Rewarding Reasoning Topology
SARL rewards reasoning topology to improve label-free RL, outperforming baselines with gains up to 44.7% on math and 34.6% on open-ended tasks while maintaining more stable training.
-
Overcoming Catastrophic Forgetting in Visual Continual Learning with Reinforcement Fine-Tuning
RaPO reduces catastrophic forgetting in visual continual learning by shaping rewards around policy drift and stabilizing advantages with cross-task exponential moving averages during reinforcement fine-tuning of multimodal models.
-
BadDLM: Backdooring Diffusion Language Models with Diverse Targets
BadDLM implants effective backdoors in diffusion language models across concept, attribute, alignment, and payload targets by exploiting denoising dynamics while preserving clean performance.
-
Step-level Denoising-time Diffusion Alignment with Multiple Objectives
MSDDA derives a closed-form optimal reverse denoising distribution for multi-objective diffusion alignment that is exactly equivalent to step-level RL fine-tuning with no approximation error.
-
Learning to Discover at Test Time
TTT-Discover applies test-time RL to set new state-of-the-art results on math inequalities, GPU kernels, algorithm contests, and single-cell denoising using an open model and public code.
-
Towards Direct Evaluation of Harness Optimizers via Priority Ranking
Priority ranking offers a low-cost direct evaluation for harness optimizers that correlates with their real multi-step optimization performance, supported by the Shor dataset of 182 scenarios.
-
SURF: Steering the Scalarization Weight to Uniformly Traverse the Pareto Front
SURF derives weight sampling rules from the arc-length CDF of the scalarization path to uniformly traverse the Pareto front in multi-objective optimization.
-
LatentUMM: Dual Latent Alignment for Unified Multimodal Models
LatentUMM proposes dual latent alignment at modality and capacity levels plus latent dynamics stabilization to reduce semantic drift and improve consistency in unified multimodal models.
-
What should post-training optimize? A test-time scaling law perspective
Tail-extrapolated estimators approximate best-of-N policy gradients from limited training rollouts by leveraging upper-tail reward statistics under structural assumptions.
-
Disentangling Intent from Role: Adversarial Self-Play for Persona-Invariant Safety Alignment
PIA achieves lower attack success rates on persona-based jailbreaks via self-play co-evolution of attacks (PLE) and defenses (PICL) that structurally decouples safety from persona context using unilateral KL-divergence.
-
Nucleus-Image: Sparse MoE for Image Generation
A 17B-parameter sparse MoE diffusion transformer activates 2B parameters per pass and reaches competitive quality on image generation benchmarks without post-training.
-
Kling-Omni Technical Report
Kling-Omni is a unified multimodal generative system that produces cinematic videos from diverse inputs by integrating generation, editing, and intelligent reasoning in a single end-to-end model.
-
Towards Continual Expansion of Data Coverage: Automatic Text-guided Edge-case Synthesis
Automated LLM-based prompt engineering for text-to-image edge-case synthesis improves object detection robustness on the FishEye8K benchmark over naive augmentation and manual prompts.
-
A Ridge Too Far: Correcting Over-Shrinkage via Negative Regularization
Negative-capable ridge regression uses controlled negative regularization as anti-shrinkage to increase effective complexity along weak eigendirections and mitigate underfitting in small-data regression.
-
PhyWorld: Physics-Faithful World Model for Video Generation
PhyWorld improves temporal consistency and physical plausibility in video world models via flow matching fine-tuning followed by DPO on physics preference pairs, with reported gains on VBench and a custom physical-faithfulness benchmark.
-
Reward Hacking in the Era of Large Models: Mechanisms, Emergent Misalignment, Challenges
The paper introduces the Proxy Compression Hypothesis as a unifying framework explaining reward hacking in RLHF as an emergent result of compressing high-dimensional human objectives into proxy reward signals under optimization pressure.
-
Z-Image: An Efficient Image Generation Foundation Model with Single-Stream Diffusion Transformer
Z-Image is an efficient 6B-parameter foundation model for image generation that rivals larger commercial systems in photorealism and bilingual text rendering through a new single-stream diffusion transformer and streamlined training.
-
Agents Should Replace Narrow Predictive AI as the Orchestrator in 6G AI-RAN
Position paper proposes replacing fragmented narrow AI models with LLMs as the cognitive orchestrator in the RAN Intelligent Controller for Level 5 autonomous 6G networks.