Attacks on Approximate Caches in Text-to-Image Diffusion Models
Bibliographic record
Abstract
The code artifact of Attacks on Approximate Caches in Text-to-Image Diffusion Models in Usenix Sec'26 Cycle 1. Three Attacks: Convert Channel Prompt Stealing Poison Attack The embedding-2-prompt reversion model used in the experiment, named "coco_prefix-049.pt", is uploaded to `https://doi.org/10.5281/zenodo.17957900`, trained by using diffutiondb dataset. You can also train your new one by using the training scripts provided in this repo. The logo insertion at embedding space model used in this experiment is "poison_attack/poison_emb/sampled_db/clip_phrase_model.pt", trained by self-constructed dataset. The dataset construction code is in "poison_attack/poison_emb/convert_data_format.py", and the training script is "poison_attack/poison_emb/logo_insertion_model.py". You can train your new model by using the training scripts. The embedding to prompt with logo model used in this experiment is "coco-prefix_latest.pt", also uploaded to `https://doi.org/10.5281/zenodo.17957900`. The training script is "poison_attack/poison_emb/recover_prompt_with_logo_model.py", you can train your own model with this script.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.003 |
| Science and technology studies | 0.003 | 0.000 |
| Scholarly communication | 0.003 | 0.001 |
| Open science | 0.003 | 0.004 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.009 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; both teacher heads agree on what is shown here.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".