Pith Number

pith:XAFXD6OC

pith:2026:XAFXD6OC723QNOKBGIFOAU5GOD

not attested not anchored not stored refs resolved

MPU: Towards Secure and Privacy-Preserving Knowledge Unlearning for Large Language Models

Pengjun Xie, Tiantong Wang, Tiantong Wu, Wei Yang Bryan Lim, Xinyu Yan, Yurong Hao

MPU lets clients unlearn LLM knowledge without exposing the forget set or the original model parameters.

arxiv:2602.23798 v2 · 2026-02-27 · cs.LG · cs.AI · cs.CR · cs.DC

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

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{XAFXD6OC723QNOKBGIFOAU5GOD}

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

MPU achieves comparable unlearning performance to noise-free baselines, with most algorithms' average degradation well below 1% up to 10% noise, and can even outperform the noise-free baseline for some algorithms under 1% noise.

C2weakest assumption

That the harmonic denoising procedure in Post-Process sufficiently removes perturbation effects without introducing new biases or security vulnerabilities that would be visible only under stronger adversarial analysis than the reported experiments.

C3one line summary

MPU is a framework that achieves privacy-preserving unlearning for LLMs by distributing perturbed model copies for local client-side unlearning followed by server-side aggregation with harmonic denoising.

References

18 extracted · 18 resolved · 3 Pith anchors

[1] On the properties of neural machine translation: Encoder-decoder approaches 2025 · doi:10.3115/v1/w14-4012

[2] OpenUnlearning: Accelerating LLM unlearning via unified benchmarking of methods and metrics

[3] Chongyang Gao, Lixu Wang, Kaize Ding, Chenkai Weng, Xiao Wang, and Qi Zhu

[4] The Llama 3 Herd of Models · arXiv:2407.21783

[5] Unlearning as multi-task optimization: A normalized gradient difference approach with an adaptive learning rate 2025

Formal links

2 machine-checked theorem links

Receipt and verification

First computed	2026-05-17T23:39:04.403706Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

b80b71f9c2feb706b941320ae053a670c3bbcb6d120330c6e317b442921c0459

Aliases

arxiv: 2602.23798 · arxiv_version: 2602.23798v2 · doi: 10.48550/arxiv.2602.23798 · pith_short_12: XAFXD6OC723Q · pith_short_16: XAFXD6OC723QNOKB · pith_short_8: XAFXD6OC

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/XAFXD6OC723QNOKBGIFOAU5GOD \
  | 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: b80b71f9c2feb706b941320ae053a670c3bbcb6d120330c6e317b442921c0459

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "f9c0f57d2a30fd7574e44e9d050838ef4227a0fb54b3f7d3ad059941a6ea61e5",
    "cross_cats_sorted": [
      "cs.AI",
      "cs.CR",
      "cs.DC"
    ],
    "license": "http://creativecommons.org/licenses/by-nc-sa/4.0/",
    "primary_cat": "cs.LG",
    "submitted_at": "2026-02-27T08:39:36Z",
    "title_canon_sha256": "c657fbc33a0ba19806d72a279069fb0dc584a7ecc66d0c31ab2cdeb881686782"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2602.23798",
    "kind": "arxiv",
    "version": 2
  }
}