Everyone's trying to make AI safe by filtering outputs. We removed the outputs.
0% jailbreak rate. Not by alignment. By architecture.
SENTINEL - provable AI safety, live on Base. Groth16 verified. Open to audit.
0 replies
0 recasts
0 reactions
Top casts
Everyone's trying to make AI safe by filtering outputs. We removed the outputs.
0% jailbreak rate. Not by alignment. By architecture.
SENTINEL - provable AI safety, live on Base. Groth16 verified. Open to audit.