Methods › General › Attention › Attention Sinks
Attention Sinks
Introduced by Guangxuan Xiao et al. in Efficient Streaming Language Models with Attention Sinks
archive 2025-07-28 Description, source and code snippet are the archive's method entry.
The archive carries only a placeholder description for this method.
In Syntology name not resolved (unknown_name). Other entries in Syntology's graph carry this name; the lookup's note below says what that means for this name.
From the lookup's note on this name: Nothing resolved to this name, but 2 entries in Syntology's graph carry it; the most cited is the paper “Attention Sinks: A 'Catch, Tag, Release' Mechanism for Embeddings” (arXiv:2502.00919, 13 citations). None of them is well cited, so for any better known sense of the name this is still a gap in coverage; they are listed so you can judge, not as a substitute. For arXiv:2502.00919: ask the lookup about it.
Near matches the lookup lists, each a different name: SS-Attention; TF-ATTENTION; VC-Attention; TC-ATTENTION; EL-Attention; Attention-MoA; DFC attention; max-attention.
Syntology's have check for this name (free, it answers coverage only), read 2026-10-05; ask it again live. A level is the served verification level of that one answer; nothing here is summed across methods.
Papers archive 2025-07-28
archive papers tagged: 13 · with a code link: 7 · where Syntology ran a sample: 5 (5 with a run with no instrument failure, 0 where every run was a failure of Syntology's instrument) Syntology
Show: all tagged papersonly where code ran (5 of 13 tagged: 5 with a run with no instrument failure, 0 where every run was a failure of Syntology's instrument)
Page 1 of 1: papers 1 to 13 of 13, newest first by the archive's date (ties by slug), in archive order.
Repository counts are the archive's code-links table. A Syntology line states what Syntology ran from that paper's harvested code, as “N ran (of which C constructed an object rather than computing a result; K with no instrument failure: H honoured, V violated, P with no contract checked; I where Syntology's instrument failed) · U unverified”; the instrument figure counts failures of Syntology's instrument, not of the code. It is per sample and not a correctness claim. When the archive marks a repository official for the paper, the line starts with that repository's state (the archive's flag, not a verdict on who wrote the code; “community repositories only” when every sample that ran came from a community repository, “official: no sample here; runs from other or unrecorded repositories” when some came from a repository the paper names or has in its text, or from none recorded); hover it for the repositories the samples that ran came from.
-
MARS-Bench: A Multi-turn Athletic Real-world Scenario Benchmark for Dialogue Evaluation 27 May 2025 · 0 repositories · arXiv:2505.23810
-
Analysis of Attention in Video Diffusion Transformers 14 Apr 2025 · 0 repositories · arXiv:2504.10317
-
Why do LLMs attend to the first token? 3 Apr 2025 · 1 repository · arXiv:2504.02732Syntology 5 ran (of which 0 constructed an object rather than computing a result; 5 with no instrument failure: 0 honoured, 0 violated, 5 with no contract checked; 0 where Syntology's instrument failed) · 2 unverified (of 7 harvested samples)
-
Interpreting the Repeated Token Phenomenon in Large Language Models 11 Mar 2025 · 1 repository · arXiv:2503.08908
-
Attention Sinks and Outlier Features: A 'Catch, Tag, and Release' Mechanism for Embeddings 2 Feb 2025 · 0 repositories · arXiv:2502.00919
-
Task-KV: Task-aware KV Cache Optimization via Semantic Differentiation of Attention Heads 25 Jan 2025 · 0 repositories · arXiv:2501.15113
-
Attention Entropy is a Key Factor: An Analysis of Parallel Context Encoding with Full-attention-based Pre-trained Language Models 21 Dec 2024 · 0 repositories · arXiv:2412.16545
-
Seeing Clearly by Layer Two: Enhancing Attention Heads to Alleviate Hallucination in LVLMs 15 Nov 2024 · 0 repositories · arXiv:2411.09968
-
Value Residual Learning For Alleviating Attention Concentration In Transformers 23 Oct 2024 · 1 repository · arXiv:2410.17897Syntology official (archive's flag): 9 ran · 9 ran (of which 0 constructed an object rather than computing a result; 4 with no instrument failure: 0 honoured, 0 violated, 4 with no contract checked; 5 where Syntology's instrument failed) · 3 unverified (of 12 harvested samples) · 1 pointer-only (licence)
-
When Attention Sink Emerges in Language Models: An Empirical View 14 Oct 2024 · 1 repository · arXiv:2410.10781Syntology official (archive's flag): 3 ran · 3 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 2 honoured, 0 violated, 0 with no contract checked; 1 where Syntology's instrument failed) · 0 unverified (of 3 harvested samples)
-
Does RoBERTa Perform Better than BERT in Continual Learning: An Attention Sink Perspective 8 Oct 2024 · 1 repository · arXiv:2410.05648
-
Unveiling and Harnessing Hidden Attention Sinks: Enhancing Large Language Models without Training through Attention Calibration 22 Jun 2024 · 1 repository · arXiv:2406.15765Syntology official (archive's flag): 5 ran · 5 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 4 where Syntology's instrument failed) · 8 unverified (of 13 harvested samples) · 13 pointer-only (licence)
-
Efficient Streaming Language Models with Attention Sinks 29 Sep 2023 · 6 repositories · arXiv:2309.17453Syntology official (archive's flag): 6 ran · 8 ran (of which 4 constructed an object rather than computing a result; 4 with no instrument failure: 0 honoured, 0 violated, 4 with no contract checked; 4 where Syntology's instrument failed) · 3 unverified (of 11 harvested samples) · 2 pointer-only (licence)
Tasks archive 2025-07-28
11 tasks the archive attaches to papers tagged with this method, by distinct papers. A task without a page in the catalog is plain text.
| Task | Papers |
|---|---|
| Continual Learning | 1 |
| Decoder | 1 |
| Dialogue Evaluation | 1 |
| Diversity | 1 |
| Hallucination | 1 |
| Language Modeling | 1 |
| Language Modelling | 1 |
| Model Compression | 1 |
| Quantization | 1 |
| TAG | 1 |
| Video Editing | 1 |
Usage over time archive 2025-07-28
Components: the archive holds no method-to-method composition, so PwC's Components table cannot be rebuilt; the Papers list carries no Results column for the same reason (the archive does not join its leaderboard rows to method tags).
Categories archive 2025-07-28
Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections