Set diffusion factorizes likelihood over arbitrary token sets and uses a set-causal diffusion architecture to support KV caching and any-order decoding, yielding improved speed-quality tradeoffs versus prior diffusion LMs.
hub
InProceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 2 (Short Papers)
17 Pith papers cite this work, alongside 68 external citations. Polarity classification is still indexing.
17
Pith papers citing it
68
external citations · OpenAlex
hub tools
citation-role summary
other 1
citation-polarity summary
roles
other 1polarities
unclear 1representative citing papers
CalBrief is a new diagnostic benchmark showing that explicit four-way strength calibration makes LLMs over-conservative mainly due to label-space expansion, while structured organization improves gap reasoning.
Evolutionary trees from LLM weights recover ground-truth training topologies and identify key datasets and layers through phenotypic analysis.
A learned residual accounting method with retrieved-token subtraction improves over pure Top-K selection at 1% exact-support budgets on long-context benchmarks for frozen Llama models.
TokenWeave achieves up to 1.28x lower latency and 1.19x higher throughput for distributed LLM inference by enabling compute-communication overlap at small token counts via a fused AllReduce-RMSNorm kernel that uses only 2-8 SMs.
A 1.88-million-article biomedical summarization dataset is released and quality-aware selection of training data based on abstract alignment outperforms random sampling on factuality metrics.
BlockGen enables flexible blockwise diffusion modeling with mixed block sizes and ARPC sampling, finding uniform diffusion outperforms masked under ancestral sampling in few-step regimes while the gap reverses with ARPC at high NFE.
DPR-BAG generates biomedical abstracts from full texts via BOMRC decomposition, parallel LLM summarization, and refinement, showing higher abstractive novelty than baselines while preserving factual consistency on a 46k-article PMC dataset.
Bell-shaped time sampling — drawing the corruption level near t=0.5 — accelerates masked diffusion language model training by up to ~4× without changing the final loss.
Training transformers with KV sparsification during continued pretraining produces representations that admit better post-hoc KV cache compression, improving quality under memory budgets for long-context tasks.
Uptraining multi-head transformer checkpoints to grouped-query attention models achieves near multi-head quality at multi-query inference speeds using 5% additional compute.
InfoKV is an entropy-aware KV cache compression framework that integrates token-level predictive uncertainty and layer-wise representation evolution with attention scores to improve performance on long-context reasoning benchmarks.
AlignedServe uses prefix-aware batching, large CPU in-flight request pools, batch scheduling, and GPU-to-GPU KV prefetching to raise decoding throughput up to 1.98x and cut latency up to 7.4x versus prior serving systems.
A hybrid graph-based training-free framework for LLM context compression matches strong baselines and shows larger gains on long-document benchmarks.
ShadowNPU presents shadowAttn, a co-designed sparse attention system that uses NPU pilot compute and techniques like graph bucketing and per-head sparsity to minimize CPU/GPU fallback during on-device LLM inference while maintaining accuracy.
Empirical study across seven language models finds concepts represented primarily in angular structure of activations while norm affects steering stability, recommending separate angular and radial parameterization over single additive coefficients.
A tree-of-thoughts inspired hybrid extractive-abstractive LLM prompt yields better legal case judgment summaries than standard extractive or abstractive prompts.
citing papers explorer
-
Set Diffusion: Interpolating Token Orderings Between Autoregression and Diffusion for Fast and Flexible Decoding
Set diffusion factorizes likelihood over arbitrary token sets and uses a set-causal diffusion architecture to support KV caching and any-order decoding, yielding improved speed-quality tradeoffs versus prior diffusion LMs.
-
CalBrief: A Pilot Diagnostic Benchmark for Evidence-Calibrated Scientific Briefing with Large Language Models
CalBrief is a new diagnostic benchmark showing that explicit four-way strength calibration makes LLMs over-conservative mainly due to label-space expansion, while structured organization improves gap reasoning.
-
Analysis and Explainability of LLMs Via Evolutionary Methods
Evolutionary trees from LLM weights recover ground-truth training topologies and identify key datasets and layers through phenotypic analysis.
-
Residual-Mass Accounting for Partial-KV Decoding
A learned residual accounting method with retrieved-token subtraction improves over pure Top-K selection at 1% exact-support budgets on long-context benchmarks for frozen Llama models.
-
TokenWeave: Efficient Compute-Communication Overlap for Distributed LLM Inference
TokenWeave achieves up to 1.28x lower latency and 1.19x higher throughput for distributed LLM inference by enabling compute-communication overlap at small token counts via a fused AllReduce-RMSNorm kernel that uses only 2-8 SMs.
-
Less is More: Quality-Aware Training Data Selection for Scientific Summarization
A 1.88-million-article biomedical summarization dataset is released and quality-aware selection of training data based on abstract alignment outperforms random sampling on factuality metrics.
-
BlockGen: Flexible Blockwise Sequence Modeling with Hybrid Samplers
BlockGen enables flexible blockwise diffusion modeling with mixed block sizes and ARPC sampling, finding uniform diffusion outperforms masked under ancestral sampling in few-step regimes while the gap reverses with ARPC at high NFE.
-
Divide-Prompt-Refine: a Training-Free, Structure-Aware Framework for Biomedical Abstract Generation
DPR-BAG generates biomedical abstracts from full texts via BOMRC decomposition, parallel LLM summarization, and refinement, showing higher abstractive novelty than baselines while preserving factual consistency on a 46k-article PMC dataset.
-
Understanding and Accelerating the Training of Masked Diffusion Language Models
Bell-shaped time sampling — drawing the corruption level near t=0.5 — accelerates masked diffusion language model training by up to ~4× without changing the final loss.
-
Training Transformers for KV Cache Compressibility
Training transformers with KV sparsification during continued pretraining produces representations that admit better post-hoc KV cache compression, improving quality under memory budgets for long-context tasks.
-
GQA: Training Generalized Multi-Query Transformer Models from Multi-Head Checkpoints
Uptraining multi-head transformer checkpoints to grouped-query attention models achieves near multi-head quality at multi-query inference speeds using 5% additional compute.
-
Information-Aware KV Cache Compression for Long Reasoning
InfoKV is an entropy-aware KV cache compression framework that integrates token-level predictive uncertainty and layer-wise representation evolution with attention scores to improve performance on long-context reasoning benchmarks.
-
AlignedServe: Orchestrating Prefix-aware Batching to Build a High-throughput and Computing-efficient LLM Serving System
AlignedServe uses prefix-aware batching, large CPU in-flight request pools, batch scheduling, and GPU-to-GPU KV prefetching to raise decoding throughput up to 1.98x and cut latency up to 7.4x versus prior serving systems.
-
From Similarity to Structure: Training-free LLM Context Compression with Hybrid Graph Priors
A hybrid graph-based training-free framework for LLM context compression matches strong baselines and shows larger gains on long-document benchmarks.
-
ShadowNPU: System and Algorithm Co-design for NPU-Centric On-Device LLM Inference
ShadowNPU presents shadowAttn, a co-designed sparse attention system that uses NPU pilot compute and techniques like graph bucketing and per-head sparsity to minimize CPU/GPU fallback during on-device LLM inference while maintaining accuracy.
-
A Geometric Account of Activation Steering through Angle-Norm Decomposition
Empirical study across seven language models finds concepts represented primarily in angular structure of activations while norm affects steering stability, recommending separate angular and radial parameterization over single additive coefficients.
-
A Tree-of-Thoughts Inspired Hybrid Approach for Legal Case Judgement Summarization using LLMs
A tree-of-thoughts inspired hybrid extractive-abstractive LLM prompt yields better legal case judgment summaries than standard extractive or abstractive prompts.