Adversarial Defense without Adversarial Defense: Enhancing Language Model Robustness via Instance-level Principal Component Removal.

Wang Y, Xiao C, Li Y, Middleton SE, Moubayed NA, Lin C

DOI: 10.1162/tacl.a.43
Published: 2025
Container: Transactions of the Association for Computational Linguistics
Publisher: Not recorded
Open access: yes

Credibility signals

limited evidence Score 45/100 under policy 1.0.0. This is a metadata assessment, not a judgment of the paper's conclusions.

Show all credibility signals

cautionDOI registered: No matching Crossref record was present in this response. Source: Crossref; license: CC0 metadata
cautionDOI resolves: No matching Crossref record was present in this response. Source: Crossref; license: CC0 metadata
not scoredDirectory of Open Access Journals: No matching DOAJ record was present in this response. No allow-list match; this is not evidence of low credibility. Source: DOAJ; license: CC0
not scoredMEDLINE indexed: Not checked or no result supplied; no credibility inference made. Source: No authority result supplied; license: Unknown
not scoredOpenAlex core source: Not checked or no result supplied; no credibility inference made. Source: No authority result supplied; license: Unknown
not scoredKnown publisher allow-list: Not checked or no result supplied; no credibility inference made. Source: No authority result supplied; license: Unknown
not scoredROR affiliation: Not checked or no result supplied; no credibility inference made. Source: No authority result supplied; license: Unknown
not scoredRetraction Watch retraction: No retraction notice matched this DOI in the deployed snapshot. No matching event found; coverage may be incomplete. Source: Retraction Watch; license: CC BY 4.0
not scoredRetraction Watch expression of concern: No expression of concern notice matched this DOI in the deployed snapshot. No matching event found; coverage may be incomplete. Source: Retraction Watch; license: CC BY 4.0
not scoredRetraction Watch correction: No correction notice matched this DOI in the deployed snapshot. No matching event found; coverage may be incomplete. Source: Retraction Watch; license: CC BY 4.0
not scoredRetraction Watch reinstatement: No reinstatement notice matched this DOI in the deployed snapshot. No matching event found; coverage may be incomplete. Source: Retraction Watch; license: CC BY 4.0
supportingOpen access status: Normalized open-access status: open. Source: Normalized work metadata; license: Caller-provided; provenance license not supplied to scorer
not scoredPublication license: Not checked or no result supplied; no credibility inference made. Source: No authority result supplied; license: Unknown
not scoredPublication version: A publication version was supplied but is not scored. Source: Normalized work metadata; license: Caller-provided; provenance license not supplied to scorer
cautionMetadata completeness: 5 of 6 scored descriptive metadata groups are present; missing fields increase uncertainty. Source: Normalized work metadata; license: Caller-provided; provenance license not supplied to scorer

Cite this work

BibTeX

@article{allodium:10.1162/tacl.a.43,
  title = {Adversarial Defense without Adversarial Defense: Enhancing Language Model Robustness via Instance-level Principal Component Removal.},
  author = {Wang Y and Xiao C and Li Y and Middleton SE and Moubayed NA and Lin C},
  year = {2025},
  journal = {Transactions of the Association for Computational Linguistics},
  doi = {10.1162/tacl.a.43},
  url = {https://doi.org/10.1162/tacl.a.43}
}

RIS

TY  - JOUR
TI  - Adversarial Defense without Adversarial Defense: Enhancing Language Model Robustness via Instance-level Principal Component Removal.
AU  - Wang Y
AU  - Xiao C
AU  - Li Y
AU  - Middleton SE
AU  - Moubayed NA
AU  - Lin C
PY  - 2025
JO  - Transactions of the Association for Computational Linguistics
DO  - 10.1162/tacl.a.43
UR  - https://doi.org/10.1162/tacl.a.43
ER  -

APA

Y, W., C, X., Y, L., SE, M., NA, M., & C, L. (2025). Adversarial Defense without Adversarial Defense: Enhancing Language Model Robustness via Instance-level Principal Component Removal.. Transactions of the Association for Computational Linguistics. https://doi.org/10.1162/tacl.a.43

Source records

pubmed · retrieved 2026-09-26T10:48:26.421Z