Pith Number

pith:RADVO4ZG

pith:2025:RADVO4ZGCQO7VFKWNOFW3QRZEV

not attested not anchored not stored refs pending

Diffusion-Inspired Masked Fine-Tuning for Knowledge Injection in Autoregressive LLMs

Ely Hahami, Haim Sompolinsky, Jingxuan Fan, Xu Pan, Ziqian Xie

Masked fine-tuning lets autoregressive LLMs absorb new facts without paraphrases and without reversal-curse failures.

arxiv:2510.09885 v6 · 2025-10-10 · 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{RADVO4ZGCQO7VFKWNOFW3QRZEV}

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

The masked fine-tuning for arLLMs substantially improves the efficacy of knowledge injection, i.e. no paraphrase needed and resistant to the reversal curse, closing the gap between arLLMs and dLLMs.

C2weakest assumption

That the demasking objective alone induces the observed knowledge-injection advantage in arLLMs independent of diffusion-specific architecture details, and that the controlled experiments isolate this effect without confounding differences in model scale, data distribution, or masking implementation.

C3one line summary

Masked fine-tuning enables autoregressive LLMs to inject new factual knowledge without paraphrases and with reversal-curse resistance, matching diffusion LLM advantages on QA tasks.

Cited by

2 papers in Pith

Knowledge Editing in Masked Diffusion Language Models

The Illusion of Latent Generalization: Bi-directionality and the Reversal Curse

Receipt and verification

First computed	2026-06-11T01:09:19.890707Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

8807577326141dfa95566b8b6dc2392562d3520316a17d2ff70b9bd2e7b769c0

Aliases

arxiv: 2510.09885 · arxiv_version: 2510.09885v6 · doi: 10.48550/arxiv.2510.09885 · pith_short_12: RADVO4ZGCQO7 · pith_short_16: RADVO4ZGCQO7VFKW · pith_short_8: RADVO4ZG

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/RADVO4ZGCQO7VFKWNOFW3QRZEV \
  | 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: 8807577326141dfa95566b8b6dc2392562d3520316a17d2ff70b9bd2e7b769c0

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "b54915ca9ea819f2c347239d64dc926c9b8813cb1d7b177bb46ddbe7083d99a2",
    "cross_cats_sorted": [
      "cs.AI"
    ],
    "license": "http://creativecommons.org/licenses/by-nc-sa/4.0/",
    "primary_cat": "cs.CL",
    "submitted_at": "2025-10-10T21:43:50Z",
    "title_canon_sha256": "ac8e2cde2e0a5981f7ccbd503bdf9f1b6993a0e2a55da1770bad22efc1ee93e2"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2510.09885",
    "kind": "arxiv",
    "version": 6
  }
}