Current vision-language and text-only models perform poorly on deciphering non-Unicode rare scripts, and Unicode encoding helps only for common languages, not low-resource ones.
McCoy, and Dragomir Radev
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Reasoning Over the Glyphs: Evaluation of LLM's Decipherment of Rare Scripts
Current vision-language and text-only models perform poorly on deciphering non-Unicode rare scripts, and Unicode encoding helps only for common languages, not low-resource ones.