SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2601.21709 · 2026

Why Attention Patterns Exist: A Unifying Temporal Perspective Analysis

Jianye Hao, Bin Li, Jie Wang, Mingxuan Yuan, Xing Li, Qingyue Yang, Huiling Zhen, Xialiang Tong, Yinqi Bai, Ch

arXiv · PDF · Open in the Atlas

Code that ran

2 of 5 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 2 failures of our instrument; 3 no recorded run), measured 2026-11-22; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2601.21709.json · badge: /api/ran/2601.21709.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
MIRALab-USTC/LLM-TAPPA community
Papers with Code archive
2 ran · 3 no recorded run
FunctionStatusWhere it lives
calculate_entropy Ran
ran; our contract draft was wrong, not the code
MIRALab-USTC/LLM-TAPPA/KVCache/tappa/model/llama_forward.py
code served (permissive licence) · get_code("518b8a174b0afc4e")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned Tensor, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# Create a small attention scores tensor
# The function expects a tensor that can be converted to float32
# and supports element-wise operations (multiplication, log, sum)
attention_scores = torch.rand(2, 4, dtype=torch.float32)

# Normalize to make valid probability distribution (sum to 1)
attention_scores = attention_scores / attention_scores.sum()

ARGS = {
    "attention_scores": attention_scores
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
compute_q_sim_torch Ran
ran; our contract draft was wrong, not the code
MIRALab-USTC/LLM-TAPPA/KVCache/tappa/model/llama_forward.py
code served (permissive licence) · get_code("2685491bc8254d65")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned float, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import math

torch.manual_seed(0)

# Create a small query tensor with shape [H, T, D]
# H=2 (batch), T=5 (time steps), D=4 (feature dimension)
q_tensor = torch.randn(2, 5, 4)

ARGS = {
    'q': q_tensor,
    'mode': 'cosine',
    'rbf_sigma': 1.0,
    'kl_eps': 1e-08
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
CakeCache Not yet run
unverified (harvested, no recorded run)
MIRALab-USTC/LLM-TAPPA/KVCache/tappa/model/llama_forward.py
code served (permissive licence) · get_code("645fbd38a3691771")
How it ran

no invocation recorded.

CakeDecodingKVCache_LayerWise Not yet run
unverified (harvested, no recorded run)
MIRALab-USTC/LLM-TAPPA/KVCache/tappa/model/llama_forward.py
code served (permissive licence) · get_code("96d4e192f7d7360e")
How it ran

no invocation recorded.

llama_attn_forward_tappa Not yet run
unverified (harvested, no recorded run)
MIRALab-USTC/LLM-TAPPA/KVCache/tappa/model/llama_forward.py
code served (permissive licence) · get_code("94b5083393355144")
How it ran

no invocation recorded.

Also present in other papers' code

Of 5 samples lifted from this paper's linked repositories, 5 carry a function identity and 0 do not. Of the 5 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 5 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Attention patterns play a crucial role in both training and inference of large language models (LLMs). Prior works have identified individual patterns such as retrieval heads, sink heads, and diagonal traces, yet these observations remain fragmented and lack a unifying explanation. To bridge this gap, we introduce Temporal Attention Pattern Predictability Analysis (TAPPA), a unifying framework that explains diverse attention patterns by analyzing their underlying mathematical formulations from a temporally continuous perspective. TAPPA both deepens the understanding of attention behavior and guides inference acceleration approaches. Specifically, TAPPA characterizes attention patterns as predictable patterns with clear regularities and unpredictable patterns that appear effectively random. Our analysis further reveals that this distinction can be explained by the degree of query self-similarity along the temporal dimension. Focusing on the predictable patterns, we further provide a detailed mathematical analysis of three representative cases through the joint effect of queries, keys, and Rotary Positional Embeddings (RoPE). We validate TAPPA by applying its insights to KV cache compression and LLM pruning tasks. Across these tasks, a simple metric motivated by TAPPA consistently improves performance over baseline methods. The code is available at https: //github.com/MIRALab-USTC/LLM-TAPPA.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2601.21709")
get_code_for_paper("2601.21709")
have("2601.21709")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2601.21709.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2601.21709.svg)](https://syntology.ai/paper/2601.21709)

Connect an agent — have() is free.