Unverified50% confidenceFactExact time
gg-friggin-ez claims to detect evasion techniques such as leetspeak, ASCII art, spaces inserted between characters, and romanized spelling.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/22/2026
First Seen
Valid until: 12/21/2026
Sources
Related Entities
Related Claims
UnverifiedIn ASCII rendering, dense characters such as #, @, and % are used to represent dark areas, while sparse characters such as . and , are used to represent light areas61% similarUnverifiedTencent Hunyuan open-sourced KSBench, a character set hallucination evaluation benchmark that detects the anti-hallucination ability of OCR visual language models by injecting character set corruption into academic paper images.58% similarUnverifiedAnti-AI fonts subtly distort characters visually, add noise, or alter shapes to keep text human-readable while interfering with OCR systems' and large language models' text parsing.58% similarUnverifiedIn typographic design, the combination of serif and sans-serif typefaces is referred to as 'contrast pairing,' which creates visual hierarchy through the contrast in typographic character.57% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/940595API
curl https://kongchang.com/api/v1/knowledge/claims/940595MCP
get_claim(id=940595)