Unverified50% confidenceBenchmarkExact time
On gpt-oss-20b, probing its reasoning process achieved an AUROC of 0.91.
1
Sources
50%
Confidence
Long-term
Relevance
10/30/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedIn the SHADE-Arena evaluation, the probe reached 98.8% AUC, surpassing the Opus 5.5-based text monitoring baseline.67% similarUnverifiedReachNT's urgency model achieves a ROC-AUC of 0.84 on unseen phrasing, and as part of an overall system that includes human review it catches 99.3% of dangerous work orders.67% similarUnverifiedgpt-oss-120b scored 0.625, lower than gpt-oss-20b's 0.720, indicating that a larger model does not necessarily mean a safer one.66% similarUnverifiedGaussian Splatting EBM achieves an AUROC of 0.9585 on 20 text embedding anomaly detection tasks, and kNN-density achieves 0.9584, with the two nearly tied.66% similarUnverifiedAureaCam's composition scoring range is from 0 to 100.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1115385API
curl https://kongchang.com/api/v1/knowledge/claims/1115385MCP
get_claim(id=1115385)