Pith Number

pith:MSLFK6E5

pith:2026:MSLFK6E5VWBDBPOYGV4T7O5FQC

not attested not anchored not stored refs resolved

E-PMQ: Expert-Guided Post-Merge Quantization with Merged-Weight Anchoring

Hongxia Yang, Jianmin Wu, Pengkai Wang, Shuo Cai, Wenjun Wang, Yanggan Gu, Yuanyi Wang

Expert-guided calibration with source experts and merged-weight anchoring makes post-merge quantization reliable for multi-task models.

arxiv:2605.16882 v1 · 2026-05-16 · cs.CL

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

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{MSLFK6E5VWBDBPOYGV4T7O5FQC}

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

On CLIP-ViT-B/32 eight-task merging, E-PMQ improves 4-bit GPTQ from 65.0% to 73.6% under Task Arithmetic and from 69.1% to 74.8% under TIES-Merging; on harder 20-task CLIP-ViT-L/14 it raises accuracy from 34.8% to 76.7%.

C2weakest assumption

That source expert weights remain available after merging and can be used to supply reliable output targets during layer-wise calibration without introducing distribution shift or extra bias relative to the merged model's integrated behavior.

C3one line summary

E-PMQ improves 4-bit quantization accuracy on merged models by 8-42 points across CLIP and GLUE tasks through expert-guided calibration and merged-weight anchoring.

References

49 extracted · 49 resolved · 3 Pith anchors

[1] and Mozer, Michael C

[2] and Beeman, David , title =

[3] and Schnell, Eric and Barkai, Edi , title =

[4] Journal of Modern Power Systems and Clean Energy , volume= 2026

[5] Democratizing AI through model fusion: A comprehensive review and future directions , author=. Nexus , year=

Formal links

2 machine-checked theorem links

Receipt and verification

First computed	2026-05-20T00:03:28.082762Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

649655789dad8230bdd835793fbba580b2883421a527c4d81ccb68baf1ed1a34

Aliases

arxiv: 2605.16882 · arxiv_version: 2605.16882v1 · doi: 10.48550/arxiv.2605.16882 · pith_short_12: MSLFK6E5VWBD · pith_short_16: MSLFK6E5VWBDBPOY · pith_short_8: MSLFK6E5

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/MSLFK6E5VWBDBPOYGV4T7O5FQC \
  | 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: 649655789dad8230bdd835793fbba580b2883421a527c4d81ccb68baf1ed1a34

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "0e2e05f164a53ed7efa077a733e494ebed8df9df3b3b4a8f5aa062834335d868",
    "cross_cats_sorted": [],
    "license": "http://creativecommons.org/licenses/by/4.0/",
    "primary_cat": "cs.CL",
    "submitted_at": "2026-05-16T08:44:36Z",
    "title_canon_sha256": "c572008e0482205524d43487271422336396af3c6166490a952b8e2e322b0141"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2605.16882",
    "kind": "arxiv",
    "version": 1
  }
}