Paper Citation Record · LEDGER

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance

As of 15 August 2026, this Paper Citation Record lists 49 of 49 outbound references and 0 inbound Pith citation observations for arXiv:2412.08521.

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

pith.paper-citation-record.v1
2412.08521 v2

Coverage vector

measured 49 of 49 reference resolution

Typed states for the displayed outbound observations.

Source: paper_references, paper_reference_links, observed 2026-08-11T17:51:08.506005Z

measured 49 of 49 standing notices

One-hop event checks from named stored sources.

Source: scholarly_work_events, retraction_status_cache, observed 2026-08-15T06:32:42.880941+00:00

measured 0 of 0 inbound itemization

Pith citing papers itemized under the disclosed page cap.

Source: paper_references, paper_reference_links

measured 0 of 1 external citation measurements

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

Source: cited_works

Reference resolution

49 of 49 outbound references displayed

verified exact0
verified fuzzy0
unresolved49
parse uncertain0
malformed identifier0
metadata mismatch0

External citation measurements

No source-named external measurement is stored.

Outbound references

Observation 7d8abfe4-5a66-4516-bd58-776076f1fafa · outbound

This paper cites an unresolved cited work.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance Unresolved cited work

Reference 1

Resolution: unresolved; no resolver link, observed 2026-08-11T17:51:08.250334Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 96d2b5a9-e9fd-4d71-b0d6-2e7bbf1a20ff · outbound

This paper cites GQA: Training Generalized Multi-Query Transformer Models from Multi-Head Checkpoints.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance GQA: Training Generalized Multi-Query Transformer Models from Multi-Head Checkpoints

Reference 2

Resolution: unresolved; no resolver link, observed 2026-08-11T17:51:08.257163Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation ddcc4913-44d6-4592-85f2-0414cdaf6413 · outbound

This paper cites an unresolved cited work.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance Unresolved cited work

Reference 3

Resolution: unresolved; raw_fallback, observed 2026-08-11T17:51:09.347766Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation e6dfa1e5-e403-4bc4-8686-911558456b0c · outbound

This paper cites PaLM 2 Technical Report.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance PaLM 2 Technical Report

Reference 4

Resolution: unresolved; no resolver link, observed 2026-08-11T17:51:08.269016Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 648e1480-c3f4-44d2-8352-69cb3ddd2b66 · outbound

This paper cites LongBench: A Bilingual, Multitask Benchmark for Long Context Understanding.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance LongBench: A Bilingual, Multitask Benchmark for Long Context Understanding

Reference 5

Resolution: unresolved; no resolver link, observed 2026-08-11T17:51:08.274585Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 06890a72-d0a3-4d2d-a694-70c653ec16c4 · outbound

This paper cites Token Merging: Your ViT But Faster.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance Token Merging: Your ViT But Faster

Reference 6

Resolution: unresolved; no resolver link, observed 2026-08-11T17:51:08.280635Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 79f9c36a-2d8c-4f20-9502-0a60f6c6380e · outbound

This paper cites an unresolved cited work.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance Unresolved cited work

Reference 7

Resolution: unresolved; no resolver link, observed 2026-08-11T17:51:08.286317Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 87fa1e17-1a3e-4b3f-b20a-788e1fd15099 · outbound

This paper cites an unresolved cited work.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance Unresolved cited work

Reference 8

Resolution: unresolved; raw_fallback, observed 2026-08-11T17:51:09.317505Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation fd16d861-a2fe-467c-bf3c-4126aaa9614a · outbound

This paper cites Extending Context Window of Large Language Models via Positional Interpolation.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance Extending Context Window of Large Language Models via Positional Interpolation

Reference 9

Resolution: unresolved; no resolver link, observed 2026-08-11T17:51:08.298079Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation aad3fd7a-14a9-4f8e-9d29-954b94f94dd6 · outbound

This paper cites an unresolved cited work.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance Unresolved cited work

Reference 10

Resolution: unresolved; raw_fallback, observed 2026-08-11T17:51:09.301051Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation cd1ce722-8526-4c03-9141-7a645b88aa77 · outbound

This paper cites an unresolved cited work.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance Unresolved cited work

Reference 11

Resolution: unresolved; raw_fallback, observed 2026-08-11T17:51:09.284755Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation c7250716-6ac6-4b9d-93e8-54e88cd1338f · outbound

This paper cites an unresolved cited work.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance Unresolved cited work

Reference 12

Resolution: unresolved; no resolver link, observed 2026-08-11T17:51:08.315443Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f514cee9-f073-4c75-a96a-c3f89c3eb975 · outbound

This paper cites BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding

Reference 13

Resolution: unresolved; no resolver link, observed 2026-08-11T17:51:08.320491Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 7e51b40d-4882-4c0b-afd1-7f040061503c · outbound

This paper cites an unresolved cited work.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance Unresolved cited work

Reference 14

Resolution: unresolved; raw_fallback, observed 2026-08-11T17:51:09.257951Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 2250b61d-c592-4bd5-84ac-bca59b44d98c · outbound

This paper cites The Llama 3 Herd of Models.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance The Llama 3 Herd of Models

Reference 15

Resolution: unresolved; no resolver link, observed 2026-08-11T17:51:08.331617Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 5f432c36-761f-4873-a215-207cd3fb4134 · outbound

This paper cites an unresolved cited work.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance Unresolved cited work

Reference 16

Resolution: unresolved; raw_fallback, observed 2026-08-11T17:51:09.237306Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 92a34430-efba-4c91-9a42-93d7e4b361cb · outbound

This paper cites an unresolved cited work.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance Unresolved cited work

Reference 17

Resolution: unresolved; raw_fallback, observed 2026-08-11T17:51:09.218745Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 7bdae81c-5e20-4e8e-aba7-b1758e9e5031 · outbound

This paper cites an unresolved cited work.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance Unresolved cited work

Reference 18

Resolution: unresolved; raw_fallback, observed 2026-08-11T17:51:09.202579Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 0415a5eb-688d-4d47-ac07-14f4db4a2995 · outbound

This paper cites Mistral 7B.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance Mistral 7B

Reference 19

Resolution: unresolved; no resolver link, observed 2026-08-11T17:51:08.355177Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b6929041-34bb-4d9e-bb5b-c5bb6035176d · outbound

This paper cites LLM Maybe LongLM: Self-Extend LLM Context Window Without Tuning.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance LLM Maybe LongLM: Self-Extend LLM Context Window Without Tuning

Reference 20

Resolution: unresolved; no resolver link, observed 2026-08-11T17:51:08.361162Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 98d45a12-a3fe-4b24-adfa-5d1c133d8bcf · outbound

This paper cites an unresolved cited work.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance Unresolved cited work

Reference 21

Resolution: unresolved; raw_fallback, observed 2026-08-11T17:51:09.184968Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation cfb5ec3e-30b6-4491-a21f-cb573033f36a · outbound

This paper cites an unresolved cited work.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance Unresolved cited work

Reference 22

Resolution: unresolved; raw_fallback, observed 2026-08-11T17:51:09.168507Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 31c513f3-4cf3-48cd-8d48-8d7997dcdb2d · outbound

This paper cites an unresolved cited work.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance Unresolved cited work

Reference 23

Resolution: unresolved; raw_fallback, observed 2026-08-11T17:51:09.152171Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 06a69042-d125-4b03-a23f-4b69a76875f4 · outbound

This paper cites PRANCE: Joint Token-Optimization and Structural Channel-Pruning for Adaptive ViT Inference.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance PRANCE: Joint Token-Optimization and Structural Channel-Pruning for Adaptive ViT Inference

Reference 24

Resolution: unresolved; no resolver link, observed 2026-08-11T17:51:08.381564Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 4f7795cc-2ed0-4a46-8c2a-1dce4bffcdad · outbound

This paper cites SnapKV: LLM Knows What You are Looking for Before Generation.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance SnapKV: LLM Knows What You are Looking for Before Generation

Reference 25

Resolution: unresolved; no resolver link, observed 2026-08-11T17:51:08.387060Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation da54a18f-e956-43a5-8730-e7e9c5cac780 · outbound

This paper cites an unresolved cited work.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance Unresolved cited work

Reference 26

Resolution: unresolved; no resolver link, observed 2026-08-11T17:51:08.392288Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation eb9b0875-23ec-4c91-a4b7-1ac130731331 · outbound

This paper cites an unresolved cited work.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance Unresolved cited work

Reference 27

Resolution: unresolved; raw_fallback, observed 2026-08-11T17:51:09.123394Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 750e56db-44de-44b3-ae67-8bfb0e12382f · outbound

This paper cites an unresolved cited work.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance Unresolved cited work

Reference 28

Resolution: unresolved; raw_fallback, observed 2026-08-11T17:51:09.106454Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation a7166149-8968-490b-9a02-1e9e3a6b24b4 · outbound

This paper cites an unresolved cited work.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance Unresolved cited work

Reference 29

Resolution: unresolved; raw_fallback, observed 2026-08-11T17:51:09.089781Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation d59a6dd5-b204-45bc-aed5-891e7cee3aba · outbound

This paper cites an unresolved cited work.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance Unresolved cited work

Reference 30

Resolution: unresolved; raw_fallback, observed 2026-08-11T17:51:09.073031Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 1022d7e2-5b88-408b-bd34-d3445390c7e0 · outbound

This paper cites an unresolved cited work.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance Unresolved cited work

Reference 31

Resolution: unresolved; raw_fallback, observed 2026-08-11T17:51:09.057248Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 057128ef-7131-4acb-baf3-a22031dfbcf4 · outbound

This paper cites an unresolved cited work.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance Unresolved cited work

Reference 32

Resolution: unresolved; no resolver link, observed 2026-08-11T17:51:08.422973Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 6587171a-b273-4bd6-ac27-7025baf83c5f · outbound

This paper cites Code Llama: Open Foundation Models for Code.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance Code Llama: Open Foundation Models for Code

Reference 33

Resolution: unresolved; no resolver link, observed 2026-08-11T17:51:08.427359Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c3896273-aba5-4cac-8afa-4f8d16a1ad5d · outbound

This paper cites an unresolved cited work.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance Unresolved cited work

Reference 34

Resolution: unresolved; no resolver link, observed 2026-08-11T17:51:08.431938Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d0ed6c20-ff5b-4e8d-b35d-4f114eee85f2 · outbound

This paper cites Llama 2: Open Foundation and Fine-Tuned Chat Models.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance Llama 2: Open Foundation and Fine-Tuned Chat Models

Reference 35

Resolution: unresolved; no resolver link, observed 2026-08-11T17:51:08.436394Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f8315b02-53dc-43ad-a7f6-9b4874990118 · outbound

This paper cites an unresolved cited work.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance Unresolved cited work

Reference 36

Resolution: unresolved; raw_fallback, observed 2026-08-11T17:51:09.014730Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 4c952e06-9434-4b54-bd76-87b70099c21c · outbound

This paper cites an unresolved cited work.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance Unresolved cited work

Reference 37

Resolution: unresolved; no resolver link, observed 2026-08-11T17:51:08.446046Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c1e232cd-3a42-488f-a748-b9019674365f · outbound

This paper cites an unresolved cited work.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance Unresolved cited work

Reference 38

Resolution: unresolved; raw_fallback, observed 2026-08-11T17:51:08.981446Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 793f136c-1f99-4018-b8ee-f0fd1d9f429d · outbound

This paper cites an unresolved cited work.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance Unresolved cited work

Reference 39

Resolution: unresolved; no resolver link, observed 2026-08-11T17:51:08.454911Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 7659c00b-9ca6-45c8-bbb4-775b8172210e · outbound

This paper cites an unresolved cited work.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance Unresolved cited work

Reference 40

Resolution: unresolved; raw_fallback, observed 2026-08-11T17:51:08.951819Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 90752c67-466e-432b-84d8-b74e0937bd25 · outbound

This paper cites HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answering.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answering

Reference 41

Resolution: unresolved; no resolver link, observed 2026-08-11T17:51:08.464151Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation fd2ae97d-21bf-4c05-b590-0163eec1f1de · outbound

This paper cites an unresolved cited work.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance Unresolved cited work

Reference 42

Resolution: unresolved; raw_fallback, observed 2026-08-11T17:51:08.935570Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation ffe7a221-c0b9-4f1a-be63-6440fbb5c65e · outbound

This paper cites an unresolved cited work.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance Unresolved cited work

Reference 43

Resolution: unresolved; raw_fallback, observed 2026-08-11T17:51:08.919410Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation 2a18e5a2-0321-4034-8b3d-60713fba99ef · outbound

This paper cites LLM Inference Unveiled: Survey and Roofline Model Insights.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance LLM Inference Unveiled: Survey and Roofline Model Insights

Reference 44

Resolution: unresolved; no resolver link, observed 2026-08-11T17:51:08.479462Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c9afb520-998a-4d18-9d97-23c8e7efe397 · outbound

This paper cites an unresolved cited work.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance Unresolved cited work

Reference 45

Resolution: unresolved; raw_fallback, observed 2026-08-11T17:51:08.901840Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation e18a207e-7740-4af5-9f8d-efe79154e4f6 · outbound

This paper cites an unresolved cited work.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance Unresolved cited work

Reference 46

Resolution: unresolved; raw_fallback, observed 2026-08-11T17:51:08.881838Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-15T06:32:42.880941+00:00.

Observation b65889f8-941e-4a06-8d00-7d9a27bc3d75 · outbound

This paper cites A Survey on Efficient Inference for Large Language Models.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance A Survey on Efficient Inference for Large Language Models

Reference 47

Resolution: unresolved; no resolver link, observed 2026-08-11T17:51:08.494901Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 57942327-0801-4474-afbd-eff8c5f17393 · outbound

This paper cites online" 'onlinestring :=.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance online" 'onlinestring :=

Reference 48

Resolution: unresolved; no resolver link, observed 2026-08-11T17:51:08.500106Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d40a4c62-d4e6-4a0e-aa79-afcfe4ae3459 · outbound

This paper cites write newline.

EMS: Adaptive Evict-then-Merge Strategy for Head-wise KV Cache Compression Based on Global-Local Importance write newline

Reference 49

Resolution: unresolved; no resolver link, observed 2026-08-11T17:51:08.506005Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Pith citing papers

No inbound Pith citation observations are available.