Pith. sign in

REVIEW 2 cited by

What do we learn from inverting CLIP models?

Not yet reviewed by Pith; the record is open.

Re-run · record.json Download PDF Read on arXiv ↗

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2403.02580 v1 pith:DMVOX375 submitted 2024-03-05 cs.CV cs.LG

Hamid Kazemi , Atoosa Chegini , Jonas Geiping , Soheil Feizi , Tom Goldstein This is my paper
classification cs.CVcs.LG
keywords clipmodelsimagespromptsinvertingabilityalignmentapproach
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

We employ an inversion-based approach to examine CLIP models. Our examination reveals that inverting CLIP models results in the generation of images that exhibit semantic alignment with the specified target prompts. We leverage these inverted images to gain insights into various aspects of CLIP models, such as their ability to blend concepts and inclusion of gender biases. We notably observe instances of NSFW (Not Safe For Work) images during model inversion. This phenomenon occurs even for semantically innocuous prompts, like "a beautiful landscape," as well as for prompts involving the names of celebrities.

Discussion (0). Sign in to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Inverting the Hidden: Unveiling Multimodal Privacy Leakage in Collaborative LVLM Inference

    cs.CR 2026-08 conditional novelty 6.0 of 10

    Intermediate hidden states transmitted during collaborative LVLM inference leak enough information to reconstruct input images and text with near-perfect token accuracy.

  2. TrustCLIP: Learning Private Visual Features via Adversarial Reconstruction

    cs.CV 2026-07 conditional novelty 6.0 of 10

    A lightweight adversarially trained projection degrades generative inversion of CLIP features while preserving most classification and VLM utility.

Pith tools