Pith Number

pith:T35E6CGL

pith:2026:T35E6CGLQMTLVT372TOZWGWVQW

not attested not anchored not stored refs pending

Reasoning Primitives in Hybrid and Non-Hybrid LLMs: Do Architectural Differences Yield Advantages in State-Tracking and Recall?

Florian Mai, Lucie Flek, Nicholas Kluge Corr\^ea, Shivam Rawat

Reasoning augmentation extends the difficulty range where models stay effective on tasks mixing recall and state-tracking, with hybrid architectures showing greater robustness to rising sequential dependence than pure transformers.

arxiv:2604.21454 v2 · 2026-04-23 · cs.CL · 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{T35E6CGLQMTLVT372TOZWGWVQW}

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

reasoning augmentation provides the largest overall improvement, substantially extending the range of difficulty over which models remain effective... in certain tasks, the hybrid reasoning model remains substantially more robust as sequential dependence increases. In contrast, the transformer reasoning model degrades sharply in performance as task difficulty increases beyond a given threshold.

C2weakest assumption

That the controlled tasks accurately isolate and jointly require only the recall and state-tracking primitives without confounding factors from model scale, training data, or task construction details, and that the matched Olmo3 transformer and hybrid variants differ only in the intended architectural inductive bias.

C3one line summary

Reasoning augmentation extends the difficulty range for both architectures, but hybrid models stay robust longer than transformers as sequential dependence increases in state-based recall tasks.

Receipt and verification

First computed	2026-05-27T00:04:24.614606Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

9efa4f08cb8326bacf7fd4dd9b1ad58588c2bb6c7329398a7f9835b902f7a92d

Aliases

arxiv: 2604.21454 · arxiv_version: 2604.21454v2 · doi: 10.48550/arxiv.2604.21454 · pith_short_12: T35E6CGLQMTL · pith_short_16: T35E6CGLQMTLVT37 · pith_short_8: T35E6CGL

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/T35E6CGLQMTLVT372TOZWGWVQW \
  | 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: 9efa4f08cb8326bacf7fd4dd9b1ad58588c2bb6c7329398a7f9835b902f7a92d

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "f528f1bdad47ef4f22a56ed59463f80109aa16f6411b65c0e4297c2eeadbd22e",
    "cross_cats_sorted": [
      "cs.AI"
    ],
    "license": "http://creativecommons.org/licenses/by/4.0/",
    "primary_cat": "cs.CL",
    "submitted_at": "2026-04-23T09:13:28Z",
    "title_canon_sha256": "708353e4b5cc819887184c542ba98c76c9529c339e934399ccbe367fcf8a2eb2"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2604.21454",
    "kind": "arxiv",
    "version": 2
  }
}