Unverified50% confidenceBenchmarkExact time
Even when the underlying VLA has already undergone safety fine-tuning, the safety violation rate is still as high as 52.6%.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/16/2026
First Seen
Valid until: 1/14/2027
Sources
Related Entities
Related Claims
UnverifiedIn experiments, CarryOn reduced the safety violation rate from 52.6% to 6.1%.65% similarUnverifiedIn LSCBench, the six models' functional pass rate reached as high as 91.3%, but the highest simultaneous functional and security pass rate was only 25.1%.63% similarUnverifiedClaude 5.1's cybersecurity guardrail false-positive interventions were reduced by an average of approximately 60%, and benign false blocks on basic biology and medical questions decreased by 85% compared to the previous generation.62% similarUnverified在小样本LoRA训练中，rank值设置存在偏差-方差权衡，实践中rank=4至32是常见的安全区间60% similarUnverifiedTest pass rates driven by model laziness are false positives that pollute security evaluation metrics and create a false sense of security.60% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1046195API
curl https://kongchang.com/api/v1/knowledge/claims/1046195MCP
get_claim(id=1046195)