SISS Web Agent - Behavioral Compromise: leaderboard
SISS measures changes in hidden-text recovery between native and aggressively downsampled images. That cross-scale shift is distinct from a web agent taking an attacker-directed action. Main, web-agent and defense experiments use separate protocols; conditional remaining attack success uses the publisher’s paired counts. This board reports behavioral compromise rate (%). Lower is better.
Metric: Behavioral compromise rate (%). Source: arxiv.org. Saturation forecast: Rough model projection: around 2026. 4 models tracked. Updated 14 Nov 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o (Harness: SISS WebAgent Baseline) | 1.67 | |
| 2 | Gemini 3 Flash (Harness: SISS WebAgent Baseline) | 2.33 | |
| 3 | GPT-4o (Harness: SISS WebAgent Adversarial) | 3.67 | |
| 4 | Gemini 3 Flash (Harness: SISS WebAgent Adversarial) | 4 |
Interactive version: theaggregate.ai/benchmark?slug=siss-web-agent-behavioral-compromise · How It Works · Data refreshed daily, snapshot 2026-11-16.