Skip to main content
Search

# Adversarial Self-Play

1 related articles

Subscribe to updates
  1. arXiv

    Disguised harmful posts slip past moderation no longer: PASP adds pinyin and glyph views to adversarial self-play, cutting adaptive attack success by up to 74.49%

    Homophones, unusual glyphs and emoji can hide harmful intent beneath surface text, and moderators that read only tokens often miss it. PASP builds textual, pinyin and rendered-image views of each input, then lets an attacker rewrite against defender feedback while the defender learns from an intent-invariance reward. On Qwen3.5-9B and GLM-4.6V-Flash, PASP lowers average adaptive attack success relative to supervised fine-tuning by up to 74.49%, with only small changes in general capability.

You’re all caught up