Paper Citation Record · LEDGER

SageAttention2++: A More Efficient Implementation of SageAttention2

As of 14 August 2026, this Paper Citation Record lists 21 of 21 outbound references and 8 inbound Pith citation observations for arXiv:2505.21136.

A citation records a reference. It does not transfer a finding from one paper to another.

pith.paper-citation-record.v1
2505.21136 v3

Coverage vector

measured 21 of 21 reference resolution

Typed states for the displayed outbound observations.

Source: paper_references, paper_reference_links, observed 2026-08-07T13:44:51.744247Z

measured 29 of 29 standing notices

One-hop event checks from named stored sources.

Source: scholarly_work_events, retraction_status_cache, observed 2026-08-13T06:32:02.005865+00:00

measured 8 of 8 inbound itemization

Pith citing papers itemized under the disclosed page cap.

Source: paper_references, paper_reference_links, observed 2026-08-09T18:02:12.529240Z

measured 0 of 1 external citation measurements

A source-named dated measurement, never combined with another source.

Source: arxiv_reference, observed 2026-07-02T05:56:40.575048Z

Reference resolution

21 of 21 outbound references displayed

verified exact0
verified fuzzy7
unresolved14
parse uncertain0
malformed identifier0
metadata mismatch0

External citation measurements

No source-named external measurement is stored.

Outbound references

Observation 705ad1c9-c08b-4c17-98bf-479d91723f4e · outbound

This paper cites SeerAttention: Learning Intrinsic Sparse Attention in Your LLMs.

SageAttention2++: A More Efficient Implementation of SageAttention2 SeerAttention: Learning Intrinsic Sparse Attention in Your LLMs

Reference 3

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:50.090294Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 076f52cd-5288-40de-b775-56e45d7e2862 · outbound

This paper cites Clipscore: A reference-free evaluation metric for image captioning.

SageAttention2++: A More Efficient Implementation of SageAttention2 Clipscore: A reference-free evaluation metric for image captioning

Reference 4

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T13:44:54.171901Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-13T06:32:02.005865+00:00.

Observation 48064749-9225-4a4b-8660-27e4158c5c48 · outbound

This paper cites an unresolved cited work.

SageAttention2++: A More Efficient Implementation of SageAttention2 Unresolved cited work

Reference 7

Resolution: unresolved; raw_fallback, observed 2026-08-07T13:44:53.878740Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-13T06:32:02.005865+00:00.

Observation d62c1c06-be2c-4eba-ba41-6ae6903e2f66 · outbound

This paper cites Online normalizer calculation for softmax.

SageAttention2++: A More Efficient Implementation of SageAttention2 Online normalizer calculation for softmax

Reference 8

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:50.535682Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1531e2d6-7dce-4ecc-ae04-6867bcba7cfa · outbound

This paper cites Wan: Open and Advanced Large-Scale Video Generative Models.

SageAttention2++: A More Efficient Implementation of SageAttention2 Wan: Open and Advanced Large-Scale Video Generative Models

Reference 11

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:50.781089Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d2e518b1-72f3-459e-9cb8-76a1d06b320c · outbound

This paper cites Linformer: Self-Attention with Linear Complexity.

SageAttention2++: A More Efficient Implementation of SageAttention2 Linformer: Self-Attention with Linear Complexity

Reference 12

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:50.843718Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 9681318e-8a69-4f65-b18c-2883a2e307c8 · outbound

This paper cites Infllm: Training-free long-context extrapolation for llms with an efficient context memory.

SageAttention2++: A More Efficient Implementation of SageAttention2 Infllm: Training-free long-context extrapolation for llms with an efficient context memory

Reference 14

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T13:44:53.344436Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-13T06:32:02.005865+00:00.

Observation 23d3377f-f866-4474-95ea-fd54adec6cc1 · outbound

This paper cites Gated Delta Networks: Improving Mamba2 with Delta Rule.

SageAttention2++: A More Efficient Implementation of SageAttention2 Gated Delta Networks: Improving Mamba2 with Delta Rule

Reference 15

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:51.129618Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation aed44000-4ddb-4930-b5b6-52076bcaf53c · outbound

This paper cites Sageattention2: Efficient attention with thorough outlier smoothing and per-thread int4 quantization.

SageAttention2++: A More Efficient Implementation of SageAttention2 Sageattention2: Efficient attention with thorough outlier smoothing and per-thread int4 quantization

Reference 16

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:51.195790Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2f17af3b-4db9-4100-b373-bcac586d42ff · outbound

This paper cites Open-Sora: Democratizing Efficient Video Production for All.

SageAttention2++: A More Efficient Implementation of SageAttention2 Open-Sora: Democratizing Efficient Video Production for All

Reference 17

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:51.272498Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 024f90a9-c7ed-4614-8813-e9f177f6dff6 · outbound

This paper cites Text-to-video models are evaluated using the open-sora (Zheng et al.,.

SageAttention2++: A More Efficient Implementation of SageAttention2 Text-to-video models are evaluated using the open-sora (Zheng et al.,

Reference 18

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T13:44:53.184410Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-13T06:32:02.005865+00:00.

Observation 8f9da372-dd34-4f2d-b2a3-3522af1caeb3 · outbound

This paper cites Text-to-image models are assessed on COCO annotations (Lin et al., 2014).

SageAttention2++: A More Efficient Implementation of SageAttention2 Text-to-image models are assessed on COCO annotations (Lin et al., 2014)

Reference 19

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T13:44:52.937859Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-13T06:32:02.005865+00:00.

Observation adac5e7d-e069-4e28-9ebe-67413cedfa95 · outbound

This paper cites For text-to-image models, generated images are compared with the images in three aspects: FID (Heusel et al.,.

SageAttention2++: A More Efficient Implementation of SageAttention2 For text-to-image models, generated images are compared with the images in three aspects: FID (Heusel et al.,

Reference 21

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T13:44:52.406675Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-13T06:32:02.005865+00:00.

Observation 924aa01a-0ec6-4e6b-9ed1-f5ea51a2e218 · outbound

This paper cites For text-to-video models, following Zhao et al.

SageAttention2++: A More Efficient Implementation of SageAttention2 For text-to-video models, following Zhao et al

Reference 1977

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T13:44:52.629890Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-13T06:32:02.005865+00:00.

Observation 34b89f3c-8c5e-482f-a8a0-c935734b2667 · outbound

This paper cites Lightning Attention-2: A Free Lunch for Handling Unlimited Sequence Lengths in Large Language Models.

SageAttention2++: A More Efficient Implementation of SageAttention2 Lightning Attention-2: A Free Lunch for Handling Unlimited Sequence Lengths in Large Language Models

Reference 2016

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:50.687143Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 6dfa40b6-696c-4e32-b3af-9204fd2c8a6d · outbound

This paper cites Identifying Sensitive Weights via Post-quantization Integral.

SageAttention2++: A More Efficient Implementation of SageAttention2 Identifying Sensitive Weights via Post-quantization Integral

Reference 2017

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:50.231009Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 70f8dc0e-b2bf-427d-b070-4ed0845d26fe · outbound

This paper cites HunyuanVideo: A Systematic Framework For Large Video Generative Models.

SageAttention2++: A More Efficient Implementation of SageAttention2 HunyuanVideo: A Systematic Framework For Large Video Generative Models

Reference 2020

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:50.310148Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 9817bad2-720d-496a-8af5-3299f7608d82 · outbound

This paper cites The Llama 3 Herd of Models.

SageAttention2++: A More Efficient Implementation of SageAttention2 The Llama 3 Herd of Models

Reference 2022

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:49.953630Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e310da7f-e507-47da-9b95-4f8ec4dc6b43 · outbound

This paper cites Sparse VideoGen: Accelerating Video Diffusion Transformers with Spatial-Temporal Sparsity.

SageAttention2++: A More Efficient Implementation of SageAttention2 Sparse VideoGen: Accelerating Video Diffusion Transformers with Spatial-Temporal Sparsity

Reference 2023

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:50.914597Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 7a7221e3-46cb-44b9-a16a-7104eadef4f2 · outbound

This paper cites Moa: Mixture of sparse attention for automatic large language model compression.

SageAttention2++: A More Efficient Implementation of SageAttention2 Moa: Mixture of sparse attention for automatic large language model compression

Reference 2024

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:50.025068Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 12738bb2-8241-492f-9413-ed5ff783b1ae · outbound

This paper cites Paperno, D., Kruszewski, G., Lazaridou, A., Pham, N.-Q., Bernardi, R., Pezzelle, S., Baroni, M., Boleda, G., and Fern´andez, R.

SageAttention2++: A More Efficient Implementation of SageAttention2 Paperno, D., Kruszewski, G., Lazaridou, A., Pham, N.-Q., Bernardi, R., Pezzelle, S., Baroni, M., Boleda, G., and Fern´andez, R

Reference 2025

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T13:44:53.580903Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-13T06:32:02.005865+00:00.

Pith citing papers

Observation 82ec07e4-dd73-4576-bff2-399298c3707c · inbound

Demystifying Cost-Efficiency in LLM Serving over Heterogeneous GPUs cites this paper.

Demystifying Cost-Efficiency in LLM Serving over Heterogeneous GPUs SageAttention2++: A More Efficient Implementation of SageAttention2

Reference 47

Resolution: unresolved; no resolver link, observed 2026-08-09T18:02:12.529240Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation bcc02c86-dd4b-43f9-b167-51bcd535b3f7 · inbound

Video Generation Models as World Models: Efficient Paradigms, Architectures and Algorithms cites this paper.

Video Generation Models as World Models: Efficient Paradigms, Architectures and Algorithms SageAttention2++: A More Efficient Implementation of SageAttention2

Reference 168

Resolution: verified exact; arxiv_id, observed 2026-05-14T01:38:35.811255Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-13T06:32:02.005865+00:00.

Observation 6e6122e8-db79-4e63-ae4a-0de51bacf2b4 · inbound

Speculative Decoding for Autoregressive Video Generation cites this paper.

Speculative Decoding for Autoregressive Video Generation SageAttention2++: A More Efficient Implementation of SageAttention2

Reference 24

Resolution: verified exact; arxiv_id, observed 2026-05-10T06:11:20.055714Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-13T06:32:02.005865+00:00.

Observation 97cf3a55-a57c-4d08-8abb-b37110a5a000 · inbound

P-Cast Precision in FP8 Attention: Sink-Induced Collapse and the Optimality of S=2^8 cites this paper.

P-Cast Precision in FP8 Attention: Sink-Induced Collapse and the Optimality of S=2^8 SageAttention2++: A More Efficient Implementation of SageAttention2

Reference 10

Resolution: verified exact; arxiv_id, observed 2026-07-02T05:56:40.576714Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-13T06:32:02.005865+00:00.

Observation 44ae696c-bff0-4237-82f5-d00d1d62a645 · inbound

Vidu S1: A Real-Time Interactive Video Generation Model cites this paper.

Vidu S1: A Real-Time Interactive Video Generation Model SageAttention2++: A More Efficient Implementation of SageAttention2

Reference 43

Resolution: unresolved; no resolver link, observed 2026-07-12T04:46:57.581402Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c63f6e16-8a37-4f22-ad1a-41449ec5fce3 · inbound

Vidu S1: A Real-Time Interactive Video Generation Model cites this paper.

Vidu S1: A Real-Time Interactive Video Generation Model SageAttention2++: A More Efficient Implementation of SageAttention2

Reference 43

Resolution: unresolved; no resolver link, observed 2026-08-02T08:56:31.582638Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 21cf43e3-b7df-4e17-9fbe-34d90684c7ad · inbound

MXAttention: Data-Free Optimal Scaling and Pre-Normalization Quantization for MXFP4 Attention cites this paper.

MXAttention: Data-Free Optimal Scaling and Pre-Normalization Quantization for MXFP4 Attention SageAttention2++: A More Efficient Implementation of SageAttention2

Reference 20

Resolution: unresolved; no resolver link, observed 2026-07-31T16:32:51.991191Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation be1ed503-9d27-423b-9bbb-e1333006e8f4 · inbound

CoSA: Accelerating Long-Context Inference via Proxy-Kernel Co-Designed Sparse Attention cites this paper.

CoSA: Accelerating Long-Context Inference via Proxy-Kernel Co-Designed Sparse Attention SageAttention2++: A More Efficient Implementation of SageAttention2

Reference 44

Resolution: unresolved; no resolver link, observed 2026-08-01T02:55:51.477723Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.