Pith Number

pith:QOPZW2B2

pith:2026:QOPZW2B2YFYM2D6YOCSNSC5YUO

not attested not anchored not stored refs pending

Conditional Memory via Scalable Lookup: A New Axis of Sparsity for Large Language Models

Bingxuan Wang, Damai Dai, Dongyan Zhao, Han Zhang, Huishuai Zhang, Kezhao Huang, Qinyu Chen, Wangding Zeng, Wenfeng Liang, Xin Cheng, Xingkai Yu, Yukun Li, Zhenda Xie, Zhewen Hao

Engram introduces conditional memory as a new sparsity axis that lets large language models perform direct O(1) knowledge lookups instead of computing retrieval.

arxiv:2601.07372 v1 · 2026-01-12 · 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{QOPZW2B2YFYM2D6YOCSNSC5YUO}

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

Scaling Engram to 27B parameters achieves superior performance over a strictly iso-parameter and iso-FLOPs MoE baseline, with notable gains in reasoning (BBH +5.0, ARC-Challenge +3.7) and long-context retrieval (Multi-Query NIAH: 84.2 to 97.0).

C2weakest assumption

The U-shaped scaling law for sparsity allocation between MoE computation and Engram memory generalizes beyond the tested model sizes and tasks, and the observed mechanistic benefits (relieving early layers, freeing attention) are causally due to the memory module rather than confounding factors in the experimental setup.

C3one line summary

Engram adds conditional memory via scalable lookup to LLMs, outperforming iso-parameter MoE baselines on reasoning and long-context tasks by following a U-shaped scaling law for allocating between computation and memory.

Formal links

2 machine-checked theorem links

Cited by

29 papers in Pith

MemCoT: Test-Time Scaling through Memory-Driven Chain-of-Thought

Exact Linear Attention

Tiny-Engram: Trigger-Indexed Concept Tables for Generative Vision

Memory Grafting: Scaling Language Model Pre-training via Offline Conditional Memory

HRM-Text: Efficient Pretraining Beyond Scaling

Receipt and verification

First computed	2026-05-17T23:38:49.048549Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

839f9b683ac170cd0fd870a4d90bb8a3a891d73cca8c323b9994b23576dabaee

Aliases

arxiv: 2601.07372 · arxiv_version: 2601.07372v1 · doi: 10.48550/arxiv.2601.07372 · pith_short_12: QOPZW2B2YFYM · pith_short_16: QOPZW2B2YFYM2D6Y · pith_short_8: QOPZW2B2

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/QOPZW2B2YFYM2D6YOCSNSC5YUO \
  | 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: 839f9b683ac170cd0fd870a4d90bb8a3a891d73cca8c323b9994b23576dabaee

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "d190569143980c41c63739ea5ab93edafb302492a5bf8dd424afdb714a6040b2",
    "cross_cats_sorted": [
      "cs.AI"
    ],
    "license": "http://creativecommons.org/licenses/by/4.0/",
    "primary_cat": "cs.CL",
    "submitted_at": "2026-01-12T09:54:49Z",
    "title_canon_sha256": "f356246ba4b44007608f772ef51593fefac15a455acd4eda5c5b20e0d67de9a2"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2601.07372",
    "kind": "arxiv",
    "version": 1
  }
}