protectai
/

deberta-v3-base-prompt-injection-v2

Text Classification

prompt-injection

Generated from Trainer

text-embeddings-inference

Model card Files Files and versions

Resources

View closed (3)

Benchmark on AgentDojo: flags 27/27 attacks alone, 23% once embedded in tool output

#9 opened 1 day ago by

A maintained, drop-in successor (same SAFE/INJECTION labels, better on benchmarks)

#8 opened 1 day ago by

Eval dataset? Very poor performance.

#7 opened 11 months ago by

using model with rust-bert

#5 opened almost 2 years ago by

Retraining on new prompt injections?

#4 opened over 2 years ago by

Hi, why i get bad results with your model ?

#2 opened over 2 years ago by