Pith Number

pith:NCYDSJED

pith:2026:NCYDSJEDP3VZTSOIJV3RCU3TCW

not attested not anchored not stored refs resolved

L2R: Low-Rank and Lipschitz-Controlled Routing for Mixture-of-Experts

Guang Li, Miki Haseyama, Minghao Yang, Ren Togo, Takahiro Ogawa

Projecting mixture-of-experts routing to a low-rank latent space with saturated inner-product scoring yields smoother geometry and stronger expert specialization.

arxiv:2601.21349 v2 · 2026-01-29 · cs.LG · cs.AI

Open paper page JSON Open Graph Bundle Merged state Verified badge What is a Pith Number?

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{NCYDSJEDP3VZTSOIJV3RCU3TCW}

Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more · Embed verified badge

Record completeness

1 Bitcoin timestamp

2 Internet Archive

3 Author claim open · sign in to claim

4 Citations open

5 Replications open

✓ Portable graph bundle live · download bundle · merged state

The bundle contains the canonical record plus signed events. A mirror can host it anywhere and recompute the same current state with the deterministic merge algorithm.

Claims

C1strongest claim

L2R consistently improves routing geometry, expert discrimination, and overall model performance on OLMoE-based language models and ImageNet vision MoE settings.

C2weakest assumption

That projecting to a low-rank latent space plus saturated inner-product scoring preserves sufficient expressiveness while only improving stability, without hidden costs in specialization or generalization.

C3one line summary

L2R improves MoE performance by routing in a low-rank space with Lipschitz-controlled saturated inner-product scoring and multi-anchor mechanisms.

References

19 extracted · 19 resolved · 10 Pith anchors

[1] BoolQ: Exploring the Surprising Difficulty of Natural Yes/No Questions 1905 · arXiv:1905.10044

[2] Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge · arXiv:1803.05457

[3] DeepSeekMoE: Towards Ultimate Expert Specialization in Mixture-of-Experts Language Models · arXiv:2401.06066

[4] Toy Models of Superposition · arXiv:2209.10652

[5] Mixtral of Experts · arXiv:2401.04088

Formal links

2 machine-checked theorem links

Receipt and verification

First computed	2026-05-17T23:39:16.545231Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

68b03924837eeb99c9c84d7711537315a57b7d7505fac923aad55b81a26547fe

Aliases

arxiv: 2601.21349 · arxiv_version: 2601.21349v2 · doi: 10.48550/arxiv.2601.21349 · pith_short_12: NCYDSJEDP3VZ · pith_short_16: NCYDSJEDP3VZTSOI · pith_short_8: NCYDSJED

Agent API

Resolver JSON Graph JSON Events JSON Schema Signing key

Verify this Pith Number yourself

curl -sH 'Accept: application/ld+json' https://pith.science/pith/NCYDSJEDP3VZTSOIJV3RCU3TCW \
  | 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: 68b03924837eeb99c9c84d7711537315a57b7d7505fac923aad55b81a26547fe

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "57247740a429ef8809cedf4b48fc4d8a3688f1b6917c1e7097ae9eefffd00194",
    "cross_cats_sorted": [
      "cs.AI"
    ],
    "license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
    "primary_cat": "cs.LG",
    "submitted_at": "2026-01-29T07:18:33Z",
    "title_canon_sha256": "242bf9a45d10ed4f2d12013d1c8f0965812c6006c27be8a40561b453f1875a72"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2601.21349",
    "kind": "arxiv",
    "version": 2
  }
}