File size: 2,916 Bytes
68a1ae6
 
d56f2fe
 
68a1ae6
d56f2fe
 
 
 
 
 
 
 
a8c5bdb
68a1ae6
 
d56f2fe
68a1ae6
 
d56f2fe
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
35042ba
 
 
 
 
 
d56f2fe
 
 
 
 
 
68a1ae6
 
d56f2fe
 
2602535
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
{
  "schema_version": 1,
  "title": "Reproduction: Rotary Position Encodings for Graphs",
  "emoji": "\ud83c\udfaf",
  "space_id": "ProCreations/repro-formal-problem-solving",
  "paper": {
    "arxiv_id": "2509.22259",
    "openreview_id": "trn64znfNx"
  },
  "tags": [
    "icml2026-repro",
    "paper-trn64znfNx"
  ],
  "updated_at": "2026-08-02T00:48:38+00:00",
  "root": {
    "slug": "index",
    "title": "Reproduction: Rotary Position Encodings for Graphs",
    "file": "pages/index.md",
    "children": [
      {
        "slug": "executive-summary",
        "title": "Executive summary",
        "file": "pages/executive-summary/page.md",
        "children": []
      },
      {
        "slug": "claim-0-current-anchored-claim-map",
        "title": "Claim 0: Current anchored-claim evidence map",
        "file": "pages/claim-0-current-anchored-claim-map/page.md",
        "children": []
      },
      {
        "slug": "claim-1-wave-induced-rotary-encodings-wire-rotate-tokens-based-on-the-graph-laplacian-spectrum-to-inject-structural-information-into-attention",
        "title": "Claim 1: Wave-Induced Rotary Encodings (WIRE) rotate tokens based on the graph Laplacian spectrum to inject structural information into attention.",
        "file": "pages/claim-1-wave-induced-rotary-encodings-wire-rotate-tokens-based-on-the-graph-laplacian-spectrum-to-inject-structural-information-into-attention/page.md",
        "children": []
      },
      {
        "slug": "claim-2-wire-recovers-regular-rope-on-grids-and-depends-asymptotically-on-the-graph-effective-resistance",
        "title": "Claim 2: WIRE recovers regular RoPE on grids and depends asymptotically on the graph effective resistance.",
        "file": "pages/claim-2-wire-recovers-regular-rope-on-grids-and-depends-asymptotically-on-the-graph-effective-resistance/page.md",
        "children": []
      },
      {
        "slug": "claim-3-wire-is-compatible-with-linear-attention-unlike-bias-based-relative-position-encodings",
        "title": "Claim 3: WIRE is compatible with linear attention, unlike bias-based relative position encodings.",
        "file": "pages/claim-3-wire-is-compatible-with-linear-attention-unlike-bias-based-relative-position-encodings/page.md",
        "children": []
      },
      {
        "slug": "claim-5-experimental-benchmark-improvements",
        "title": "Claim 5: Experiments report WIRE improvements on synthetic monochromatic-subgraph and shortest-path tasks, point-cloud transformers, and efficient graph transformer benchmarks (Table 1, Table 2, Table 3, Table 4).",
        "file": "pages/claim-5-experimental-benchmark-improvements/page.md",
        "children": []
      },
      {
        "slug": "conclusion",
        "title": "Conclusion",
        "file": "pages/conclusion/page.md",
        "children": []
      }
    ]
  },
  "agent_view_tokens": 8000,
  "revision": "1784412310123238000"
}