[KongchangAI]
BenchmarkTensor Trust

Tensor Trust

A benchmark dataset for evaluating prompt injection defense capability, containing adversarial prompts and defense tasks.

Timeline (last 90 days)

Nov 10

The audit benchmark covers NLI (e-SNLI), adversarial syntactic shortcuts (HANS), synthetic clinical triage (N=900), prompt injection defense (Tensor Trust, N=250), and an 86-case logic-gate trigger suite.

Unverified50%
Nov 10

In the prompt-injection defense task (N=250), System 1 models with 4B parameters or more achieve an AUROC of 0.988 to 0.998, a recall of up to 95.2%, and a 0.0% false positive rate at 48 to 60 milliseconds of latency.

Unverified50%

All Facts (2)

Source Articles