Paper Citation Record · LEDGER

FLASH-D: FlashAttention with Hidden Softmax Division

As of 8 August 2026, this Paper Citation Record lists 34 of 34 outbound references and 1 inbound Pith citation observation for arXiv:2505.14201.

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

pith.paper-citation-record.v1
2505.14201 v1

Coverage vector

measured 34 of 34 reference resolution

Typed states for the displayed outbound observations.

Source: paper_references, paper_reference_links, observed 2026-08-07T15:42:44.481313Z

measured 35 of 35 standing notices

One-hop event checks from named stored sources.

Source: scholarly_work_events, retraction_status_cache, observed 2026-08-08T06:32:00.761636+00:00

measured 1 of 1 inbound itemization

Pith citing papers itemized under the disclosed page cap.

Source: paper_references, paper_reference_links, observed 2026-05-10T16:10:40.858525Z

measured 0 of 1 external citation measurements

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

Source: arxiv_reference, observed 2026-05-11T09:16:00.127986Z

Reference resolution

34 of 34 outbound references displayed

verified exact2
verified fuzzy19
unresolved13
parse uncertain0
malformed identifier0
metadata mismatch0

External citation measurements

No source-named external measurement is stored.

Outbound references

Observation 349993c8-feb5-4014-a864-1cfce29a038e · outbound

This paper cites An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale.

FLASH-D: FlashAttention with Hidden Softmax Division An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale

Reference 1

Resolution: unresolved; no resolver link, observed 2026-08-07T15:42:41.396596Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation cee44612-0d29-43ac-8842-f30050447758 · outbound

This paper cites DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning.

FLASH-D: FlashAttention with Hidden Softmax Division DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning

Reference 2

Resolution: unresolved; no resolver link, observed 2026-08-07T15:42:41.486142Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 7e8e00ff-03cd-4f0f-bb7f-91a51a4f3e07 · outbound

This paper cites Attention is all you need,.

FLASH-D: FlashAttention with Hidden Softmax Division Attention is all you need,

Reference 3

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:42:48.818408Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation 7d28a434-0e31-4b99-8e28-4883668f8793 · outbound

This paper cites Longformer: The Long-Document Transformer.

FLASH-D: FlashAttention with Hidden Softmax Division Longformer: The Long-Document Transformer

Reference 4

Resolution: unresolved; no resolver link, observed 2026-08-07T15:42:41.807184Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation da553e48-38be-4b9b-9751-363622a02bb1 · outbound

This paper cites Generating Long Sequences with Sparse Transformers.

FLASH-D: FlashAttention with Hidden Softmax Division Generating Long Sequences with Sparse Transformers

Reference 5

Resolution: unresolved; no resolver link, observed 2026-08-07T15:42:41.876917Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d861182a-2346-4b5d-953d-333fccd78792 · outbound

This paper cites Transformers are rnns: Fast autoregressive transformers with linear attention,.

FLASH-D: FlashAttention with Hidden Softmax Division Transformers are rnns: Fast autoregressive transformers with linear attention,

Reference 6

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:42:48.665023Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation fa7d2e7a-5ab0-4d68-95f9-2bc161848ece · outbound

This paper cites Low-rank Attention Side-Tuning for Parameter-Efficient Fine-Tuning.

FLASH-D: FlashAttention with Hidden Softmax Division Low-rank Attention Side-Tuning for Parameter-Efficient Fine-Tuning

Reference 7

Resolution: verified exact; local_arxiv, observed 2026-08-07T15:42:44.866897Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation e0b04749-b178-4e00-9827-2ec14640240c · outbound

This paper cites A3: Accelerating attention mechanisms in neural networks with approximation,.

FLASH-D: FlashAttention with Hidden Softmax Division A3: Accelerating attention mechanisms in neural networks with approximation,

Reference 8

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:42:48.422976Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation 9ddba633-3b3e-4e9a-adbc-cea455b0f1d7 · outbound

This paper cites A 95.6-TOPS/W deep learning inference accelerator with per-vector scaled 4-bit quantization in 5 nm,.

FLASH-D: FlashAttention with Hidden Softmax Division A 95.6-TOPS/W deep learning inference accelerator with per-vector scaled 4-bit quantization in 5 nm,

Reference 9

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:42:48.215809Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation c7ee989e-6d9e-4e31-be83-90ed16390a24 · outbound

This paper cites Hardware accelerator for multi-head attention and position-wise feed-forward in the trans- former,.

FLASH-D: FlashAttention with Hidden Softmax Division Hardware accelerator for multi-head attention and position-wise feed-forward in the trans- former,

Reference 10

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:42:47.975824Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation 9f4355ff-8cb8-425e-a162-65cb5a70bd09 · outbound

This paper cites Mnnfast: a fast and scalable system architecture for memory-augmented neural networks,.

FLASH-D: FlashAttention with Hidden Softmax Division Mnnfast: a fast and scalable system architecture for memory-augmented neural networks,

Reference 11

Resolution: unresolved; no resolver link, observed 2026-08-07T15:42:42.406649Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 56ed8ec6-0c96-4456-91ee-eb4ee6ddd782 · outbound

This paper cites COSA plus: Enhanced co-operative systolic arrays for attention mechanism in transformers,.

FLASH-D: FlashAttention with Hidden Softmax Division COSA plus: Enhanced co-operative systolic arrays for attention mechanism in transformers,

Reference 12

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:42:47.793607Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation 1fb37edb-fe37-4de3-abaa-efb49c65b383 · outbound

This paper cites ELSA: Hardware-software co-design for efficient, lightweight self- attention mechanism in neural networks,.

FLASH-D: FlashAttention with Hidden Softmax Division ELSA: Hardware-software co-design for efficient, lightweight self- attention mechanism in neural networks,

Reference 13

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:42:47.470310Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation afb78e9b-4af3-4d79-a5f5-22c5fa03f436 · outbound

This paper cites TSAcc: An efficient tempo-spatial similarity aware accelerator for attention acceleration,.

FLASH-D: FlashAttention with Hidden Softmax Division TSAcc: An efficient tempo-spatial similarity aware accelerator for attention acceleration,

Reference 14

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:42:47.270387Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation 4f386474-4bba-4ac7-b3fc-f64ac1eb530f · outbound

This paper cites X-former: In-memory acceleration of transformers,.

FLASH-D: FlashAttention with Hidden Softmax Division X-former: In-memory acceleration of transformers,

Reference 15

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:42:47.146014Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation 9fb1e344-3540-4a0e-8e48-92d009a77b53 · outbound

This paper cites Flashattention: Fast and memory-efficient exact attention with IO-awareness,.

FLASH-D: FlashAttention with Hidden Softmax Division Flashattention: Fast and memory-efficient exact attention with IO-awareness,

Reference 16

Resolution: unresolved; no resolver link, observed 2026-08-07T15:42:42.843511Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f153e9dd-7686-4826-9cd3-73827174f756 · outbound

This paper cites FlashAttention-2: Faster Attention with Better Parallelism and Work Partitioning.

FLASH-D: FlashAttention with Hidden Softmax Division FlashAttention-2: Faster Attention with Better Parallelism and Work Partitioning

Reference 17

Resolution: unresolved; no resolver link, observed 2026-08-07T15:42:42.955605Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d1bdae36-d87c-4e8e-bc7c-ff74eab4a9a4 · outbound

This paper cites Self-attention Does Not Need $O(n^2)$ Memory.

FLASH-D: FlashAttention with Hidden Softmax Division Self-attention Does Not Need $O(n^2)$ Memory

Reference 18

Resolution: unresolved; no resolver link, observed 2026-08-07T15:42:43.089177Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a7f5bb64-107c-4696-89dc-2b8e594bb8b9 · outbound

This paper cites Hardware-efficient softmax approximation for self-attention networks,.

FLASH-D: FlashAttention with Hidden Softmax Division Hardware-efficient softmax approximation for self-attention networks,

Reference 19

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:42:47.014152Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation 6ede0174-f242-48c7-9086-66d271395e83 · outbound

This paper cites Language models are unsupervised multitask learners,.

FLASH-D: FlashAttention with Hidden Softmax Division Language models are unsupervised multitask learners,

Reference 20

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:42:46.816425Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation f091fd2e-ee33-4128-9ef7-e09aa398d584 · outbound

This paper cites Fastervit: Fast vision transformers with hierarchical attention,.

FLASH-D: FlashAttention with Hidden Softmax Division Fastervit: Fast vision transformers with hierarchical attention,

Reference 21

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:42:46.664766Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation 07f7228f-5b5f-473f-8d4e-b13b1718f94b · outbound

This paper cites ConSmax: Hardware-Friendly Alternative Softmax with Learnable Parameters.

FLASH-D: FlashAttention with Hidden Softmax Division ConSmax: Hardware-Friendly Alternative Softmax with Learnable Parameters

Reference 22

Resolution: verified exact; local_arxiv, observed 2026-08-07T15:42:44.687202Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation e64bf920-7b6b-4733-a4e8-43710b9e17ee · outbound

This paper cites Online normalizer calculation for softmax.

FLASH-D: FlashAttention with Hidden Softmax Division Online normalizer calculation for softmax

Reference 23

Resolution: unresolved; no resolver link, observed 2026-08-07T15:42:43.564642Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c8dd91a7-88e8-490a-9ad8-596a052082d9 · outbound

This paper cites Edge- BERT: sentence-level energy optimizations for latency-aware multi-task NLP inference,.

FLASH-D: FlashAttention with Hidden Softmax Division Edge- BERT: sentence-level energy optimizations for latency-aware multi-task NLP inference,

Reference 24

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:42:46.461527Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation c3816f74-255d-41a5-b4b9-ae90019b3b14 · outbound

This paper cites Online alignment and ad- dition in multiterm floating-point adders,.

FLASH-D: FlashAttention with Hidden Softmax Division Online alignment and ad- dition in multiterm floating-point adders,

Reference 25

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:42:46.289950Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation 6be2cb52-9b72-4d12-a75e-6b2b464b6237 · outbound

This paper cites Templatized fused vector floating-point dot product for high-level synthesis,.

FLASH-D: FlashAttention with Hidden Softmax Division Templatized fused vector floating-point dot product for high-level synthesis,

Reference 26

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:42:46.107398Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation 846fc410-41d1-4944-af98-ba7a8780ab8c · outbound

This paper cites SOLE: hardware- software co-design of softmax and layernorm for efficient transformer inference,.

FLASH-D: FlashAttention with Hidden Softmax Division SOLE: hardware- software co-design of softmax and layernorm for efficient transformer inference,

Reference 27

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:42:45.921795Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation a277cdac-02a1-4c13-bfc7-0670df9ea99c · outbound

This paper cites A pseudo-softmax function for hardware-based high speed image classification,.

FLASH-D: FlashAttention with Hidden Softmax Division A pseudo-softmax function for hardware-based high speed image classification,

Reference 28

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:42:45.672152Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation 9864db61-9e6f-453b-ad20-dd321fa8a926 · outbound

This paper cites an unresolved cited work.

FLASH-D: FlashAttention with Hidden Softmax Division Unresolved cited work

Reference 29

Resolution: unresolved; raw_fallback, observed 2026-08-07T15:42:45.444137Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation ae3bb868-7e88-46ce-9cc4-dce72a816076 · outbound

This paper cites A Study of BFLOAT16 for Deep Learning Training.

FLASH-D: FlashAttention with Hidden Softmax Division A Study of BFLOAT16 for Deep Learning Training

Reference 30

Resolution: unresolved; no resolver link, observed 2026-08-07T15:42:44.143506Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 9b71812c-8888-4890-9a49-da600fbe2d7c · outbound

This paper cites FP8 Formats for Deep Learning.

FLASH-D: FlashAttention with Hidden Softmax Division FP8 Formats for Deep Learning

Reference 31

Resolution: unresolved; no resolver link, observed 2026-08-07T15:42:44.231773Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 69bfcb20-f662-49e6-ba4c-34dd47ec368b · outbound

This paper cites llama2.c,.

FLASH-D: FlashAttention with Hidden Softmax Division llama2.c,

Reference 32

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:42:45.225376Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation ac1a776d-c7f3-4bb4-ab42-eb83152947a4 · outbound

This paper cites Transformers: State-of-the-art natural language processing,.

FLASH-D: FlashAttention with Hidden Softmax Division Transformers: State-of-the-art natural language processing,

Reference 33

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T15:42:45.046848Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation e3b17ae8-85ee-4fe3-a048-05050b577aba · outbound

This paper cites PromptBench: A Unified Library for Evaluation of Large Language Models.

FLASH-D: FlashAttention with Hidden Softmax Division PromptBench: A Unified Library for Evaluation of Large Language Models

Reference 34

Resolution: unresolved; no resolver link, observed 2026-08-07T15:42:44.481313Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Pith citing papers

Observation 639b91f1-2560-466d-88fc-4a7dfc650174 · inbound

VFA: Relieving Vector Operations in Flash Attention with Global Maximum Pre-computation cites this paper.

VFA: Relieving Vector Operations in Flash Attention with Global Maximum Pre-computation FLASH-D: FlashAttention with Hidden Softmax Division

Reference 3

Resolution: verified exact; arxiv_id, observed 2026-05-11T09:16:00.130999Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.