unicode
embedding-qwen3-1.7b-embedding_ac1_unicode_shufembedding-qwen3-1.7b-embedding_ctxt_ac1_unicode_shufembedding-qwen3-1.7b-embedding_norm-ac_unicode_shufunicode-obfuscation-finetune-0.6bgrowing-transformers-model-unicode-1-9-247membedding-qwen3-1.7b-embedding_unicode_shufembedding-qwen3-1.7b-embedding_ctxt_unicode_shufembedding-qwen3-1.7b-embedding_ctxt_norm-ac_unicode_shuf
Datasets
All datasets matching “unicode”invisible-unicode-injection
Invisible Unicode Prompt Injection
from datasets import load_dataset
ds = load_dataset("fevziegeyurtsevenler/invisible-unicode-injection")
42 examples where innocent visible text hides an instruction in invisible Unicode (Tags block U+E0000–E007F, zero-width). The model reads the payload; a human reviewer sees nothing. Decode + detect them.
Each row: visible_text (innocent), hidden_payload (decoded), full_text (with the real invisible chars), technique, detected_by (uncloak… See the full description on the dataset page: https://huggingface.co/datasets/fevziegeyurtsevenler/invisible-unicode-injection.uof-unicode-mutation-graphs
UOF Unicode Mutation Graphs v2
Synthetic Unicode mutation cases with explicit code-point, normalization, bidirectional-class, general-category, identifier, printability, and action evidence.
Each example starts from a generated root string and applies one named mutation operation. The row keeps the escaped input, code points, Unicode properties, normalization deltas, evidence list, recommended action, and structured target together.
Release contract
Generator:… See the full description on the dataset page: https://huggingface.co/datasets/NewSonnet/uof-unicode-mutation-graphs.UniCoder-Instructunicodec-fisher-trainfacebook-sinhala-unicode-hate-speechunicodec-librispeech