HERALD overlaps CPU-side sparse KV-cache selection with GPU-side block-denoising to hide offloading latency, achieving up to 2.28–2.47x decode throughput over dense GPU serving at 5–10% KV budgets.
arXiv preprint arXiv:2509.06949 , year=
19 Pith papers cite this work. Polarity classification is still indexing.
19
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
years
2026 19roles
background 1polarities
background 1representative citing papers
d-OPSD reframes on-policy self-distillation for dLLMs via suffix conditioning from self-generated answers and step-level supervision, outperforming RLVR and SFT on reasoning benchmarks with ~10% of the optimization steps.
The paper introduces Uni-E, a unified energy for DLMs that accounts for model capacity, dependency and invariance, can be computed exactly, and corrects distribution shifts from dependency and invariance.
TraFL applies trajectory flow balancing to post-train diffusion language models, preventing mode collapse and delivering consistent gains on reasoning tasks that hold under increased sampling.
RSPO interprets reward advantages as targets for relative log-ratios in dLLMs, calibrating noisy estimates to stabilize RLVR training and achieve strong gains on planning tasks with competitive math reasoning performance.
TAD improves the accuracy-parallelism trade-off in diffusion LLMs via temporal-aware self-distillation that applies hard labels to soon-to-be-decoded tokens and soft supervision to future tokens.
MemDLM embeds a simulated denoising trajectory into DLM training via bi-level optimization, creating a parametric memory that improves convergence and long-context performance even when the memory is dropped at test time.
Info-Gain Sampler improves MDM decoding by using bidirectional information gain to reduce cumulative uncertainty, outperforming greedy samplers on reasoning accuracy and creative writing tasks.
Accept-Until-Fail training improves average accepted block length in speculative decoding from 2.40 to 2.61 by limiting cross-entropy support to the drafter's first predicted failure point.
CFT + on-policy distillation + RL converts an AR reasoning re-ranker into a block-diffusion model that recovers near-AR accuracy at 2.4–3.5× decode throughput on Amazon Beauty.
VoidPadding decouples padding from termination in MDLMs via a new [VOID] token, delivering +17.84 average benchmark points and 55.7% fewer decoding steps on Dream-7B-Instruct.
AGDO improves dLLM reasoning performance by determining denoising order and emphasizing tokens based on attention-derived dependencies rather than random masking.
PAPO improves reasoning performance in diffusion LLMs by converting sparse terminal rewards into dense step-wise credit and replaying real high-uncertainty trajectories, reporting gains up to 42.2% on Countdown.
On-policy distillation from a frozen autoregressive teacher to a bidirectional student eliminates train-inference mismatch and enables data-efficient ARLM-to-DLM conversion.
GDSD reduces RL for dLLMs to likelihood-free self-distillation via a normalization-free logit-matching objective, outperforming ELBO methods with more stable training on LLaDA-8B and Dream-7B.
Fast-dDrive is a block-diffusion VLA that reports SOTA accuracy on WOD-E2E and nuScenes driving benchmarks together with 12x throughput over autoregressive baselines via section scaffolds and test-time averaging.
Proposes HT-GRPO with sketch-then-paint staged updates, prompt-conditioned importance ratios, and hierarchical credit assignment for dMLLMs, reporting gains on GenEval and DPG plus quality metrics.
DIA is a training-free method that dynamically adjusts anchor positions in diffusion LLMs to improve format compliance and accuracy on reasoning benchmarks like GSM8K and MATH.
Static checking rewards and moderate AST-based hints improve diffusion RL performance for code generation, with effectiveness varying by task difficulty across HumanEval, MBPP, and LiveCodeBench.
citing papers explorer
-
HERALD: High-Throughput Block Diffusion LLM Serving via CPU-GPU Cooperative KV Cache Retrieval
HERALD overlaps CPU-side sparse KV-cache selection with GPU-side block-denoising to hide offloading latency, achieving up to 2.28–2.47x decode throughput over dense GPU serving at 5–10% KV budgets.
-
Learning from the Self-future: On-policy Self-distillation for dLLMs
d-OPSD reframes on-policy self-distillation for dLLMs via suffix conditioning from self-generated answers and step-level supervision, outperforming RLVR and SFT on reasoning benchmarks with ~10% of the optimization steps.
-
Unified Energy for Invariant and Independent Decoding in Diffusion Language Models
The paper introduces Uni-E, a unified energy for DLMs that accounts for model capacity, dependency and invariance, can be computed exactly, and corrects distribution shifts from dependency and invariance.
-
Beyond Mode-Seeking RL: Trajectory-Balance Post-Training for Diffusion Language Models
TraFL applies trajectory flow balancing to post-train diffusion language models, preventing mode collapse and delivering consistent gains on reasoning tasks that hold under increased sampling.
-
Relative Score Policy Optimization for Diffusion Language Models
RSPO interprets reward advantages as targets for relative log-ratios in dLLMs, calibrating noisy estimates to stabilize RLVR training and achieve strong gains on planning tasks with competitive math reasoning performance.
-
TAD: Temporal-Aware Trajectory Self-Distillation for Fast and Accurate Diffusion LLM
TAD improves the accuracy-parallelism trade-off in diffusion LLMs via temporal-aware self-distillation that applies hard labels to soon-to-be-decoded tokens and soft supervision to future tokens.
-
MemDLM: Memory-Enhanced DLM Training
MemDLM embeds a simulated denoising trajectory into DLM training via bi-level optimization, creating a parametric memory that improves convergence and long-context performance even when the memory is dropped at test time.
-
Improving Sampling for Masked Diffusion Models via Information Gain
Info-Gain Sampler improves MDM decoding by using bidirectional information gain to reduce cumulative uncertainty, outperforming greedy samplers on reasoning accuracy and creative writing tasks.
-
Spec-AUF: Accept-Until-Fail Training under Train-Inference Misalignment for Masked Block Drafters
Accept-Until-Fail training improves average accepted block length in speculative decoding from 2.40 to 2.61 by limiting cross-entropy support to the drafter's first predicted failure point.
-
Diffusion-GR2: Diffusion Generative Reasoning Re-ranker
CFT + on-policy distillation + RL converts an AR reasoning re-ranker into a block-diffusion model that recovers near-AR accuracy at 2.4–3.5× decode throughput on Amazon Beauty.
-
VoidPadding: Let [VOID] Handle Padding in Masked Diffusion Language Models so that [EOS] Can Focus on Semantic Termination
VoidPadding decouples padding from termination in MDLMs via a new [VOID] token, delivering +17.84 average benchmark points and 55.7% fewer decoding steps on Dream-7B-Instruct.
-
Beyond Fully Random Masking: Attention-Guided Denoising and Optimization for Diffusion Language Models
AGDO improves dLLM reasoning performance by determining denoising order and emphasizing tokens based on attention-derived dependencies rather than random masking.
-
Back on Track: Aligning Rewards and States for Reasoning in Diffusion Large Language Models
PAPO improves reasoning performance in diffusion LLMs by converting sparse terminal rewards into dense step-wise credit and replaying real high-uncertainty trajectories, reporting gains up to 42.2% on Countdown.
-
Data-Efficient Autoregressive-to-Diffusion Language Models via On-Policy Distillation
On-policy distillation from a frozen autoregressive teacher to a bidirectional student eliminates train-inference mismatch and enables data-efficient ARLM-to-DLM conversion.
-
GDSD: Reinforcement Learning as Guided Denoiser Self-Distillation for Diffusion Language Models
GDSD reduces RL for dLLMs to likelihood-free self-distillation via a normalization-free logit-matching objective, outperforming ELBO methods with more stable training on LLaDA-8B and Dream-7B.
-
Fast-dDrive: Efficient Block-Diffusion VLM for Autonomous Driving
Fast-dDrive is a block-diffusion VLA that reports SOTA accuracy on WOD-E2E and nuScenes driving benchmarks together with 12x throughput over autoregressive baselines via section scaffolds and test-time averaging.
-
Sketch Then Paint: Hierarchical Reinforcement Learning for Diffusion Multi-Modal Large Language Models
Proposes HT-GRPO with sketch-then-paint staged updates, prompt-conditioned importance ratios, and hierarchical credit assignment for dMLLMs, reporting gains on GenEval and DPG plus quality metrics.
-
Dynamic Infilling Anchors for Format-Constrained Generation in Diffusion Large Language Models
DIA is a training-free method that dynamically adjusts anchor positions in diffusion LLMs to improve format compliance and accuracy on reasoning benchmarks like GSM8K and MATH.
-
Beyond Execution: Static-Analysis Rewards and Hint-Conditioned Diffusion RL for Code Generation
Static checking rewards and moderate AST-based hints improve diffusion RL performance for code generation, with effectiveness varying by task difficulty across HumanEval, MBPP, and LiveCodeBench.