Applying a single-turn 'crescendo' prompt that pretends earlier images were already generated raises DALL-E 3's harmful image rate from 1.3% to 18.5%, close to the uncensored Flux Schnell baseline.
Tiger-Gemma-9B-v3 model from HuggingFace - decensored Gemma 9B SPPO
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CR 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
An indicator for effectiveness of text-to-image guardrails utilizing the Single-Turn Crescendo Attack (STCA)
Applying a single-turn 'crescendo' prompt that pretends earlier images were already generated raises DALL-E 3's harmful image rate from 1.3% to 18.5%, close to the uncensored Flux Schnell baseline.