pith:TEIOMHJT
All Circuits Lead to Rome: Rethinking Functional Anisotropy in Circuit and Sheaf Discovery for LLMs
A single task in large language models can be performed by multiple structurally distinct circuits or sheaves that are all faithful, sparse, and complete.
arxiv:2605.12671 v1 · 2026-05-12 · cs.CL
Add to your LaTeX paper
\usepackage{pith}
\pithnumber{TEIOMHJTIAWZPAC2D57CFCFFGH}
Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more · Embed verified badge
Record completeness
Claims
We show that a single LLM task can instead be supported by multiple, structurally distinct circuits or sheaves that are simultaneously faithful, sparse, and complete.
That the Overlap-Aware Sheaf Repulsion penalty successfully uncovers genuinely distinct mechanisms without introducing methodological artifacts, and that the theoretical analysis of high-dimensional superposition holds under the stated mild assumptions.
LLM tasks are supported by multiple distinct circuits rather than unique mechanisms, demonstrated via Overlap-Aware Sheaf Repulsion and the Distributive Dense Circuit Hypothesis.
References
Receipt and verification
| First computed | 2026-05-18T03:09:50.224482Z |
|---|---|
| Builder | pith-number-builder-2026-05-17-v1 |
| Signature | Pith Ed25519
(pith-v1-2026-05) · public key |
| Schema | pith-number/v1.0 |
Canonical hash
9910e61d33402d97805a1f7e2288a531ce206f4307ac3f8348a0d04acbd89c17
Aliases
· · · · ·Agent API
Verify this Pith Number yourself
curl -sH 'Accept: application/ld+json' https://pith.science/pith/TEIOMHJTIAWZPAC2D57CFCFFGH \
| jq -c '.canonical_record' \
| python3 -c "import sys,json,hashlib; b=json.dumps(json.loads(sys.stdin.read()), sort_keys=True, separators=(',',':'), ensure_ascii=False).encode(); print(hashlib.sha256(b).hexdigest())"
# expect: 9910e61d33402d97805a1f7e2288a531ce206f4307ac3f8348a0d04acbd89c17
Canonical record JSON
{
"metadata": {
"abstract_canon_sha256": "b372839d2a82aa15f4353102729121b01c37f1d4d0a56b1c7fe98881c2c3946b",
"cross_cats_sorted": [],
"license": "http://creativecommons.org/licenses/by/4.0/",
"primary_cat": "cs.CL",
"submitted_at": "2026-05-12T19:21:23Z",
"title_canon_sha256": "aaf33bf90f03f84b493997633b106b7f4a97226a4afbb66ab912e25b746151bf"
},
"schema_version": "1.0",
"source": {
"id": "2605.12671",
"kind": "arxiv",
"version": 1
}
}