The Decoder· Maximilian Schreiner·· 6 h ago
Anthropic's new prototype proves which model wrote an AI response without exposing its weights
Anthropic's new prototype proves which model wrote an AI response without exposing its weights
AI summary
Anthropic has released FlashProve, a research prototype that uses cryptography to verify whether an AI response actually came from the claimed model. The system still has significant limitations.
Selection record
Threshold 76Media and individualsFirst 58Second 62
Not admittedSum of both 120 < twice the threshold 152
- Source tier
- Media and individuals; this tier's threshold is 76
- Pre-filter
- passed：Anthropic发布AI推理验证原型FlashProve
A model scores each item twice, independently, against one written standard, out of 100. An item is admitted only when the two scores add up to twice the threshold. The threshold is set per source tier.
Source: The Decoder · the-decoder.com