Unverified50% confidenceFactExact time
navhard is the more difficult test subset in NAVSIM, specifically containing interaction-dense, higher-risk scenarios, and the same model's score on navhard is usually significantly lower than on navtest.
1
Sources
50%
Confidence
Long-term
Relevance
10/15/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedTraining samples for the N2 category tend to be scarcer, and data imbalance exacerbates the model's tendency to favor the majority class N1, which is one of the important contributing factors to the asymmetric recall.59% similarUnverifiedThe NIST randomness test suite has two fundamental limitations: first, it tests statistical properties, and it is often powerless against carefully constructed pseudorandom sequences; second, it does not model the intelligence level of the observer.57% similarUnverifiedA common limitation of red-team testing, capability evaluations, and alignment checks is that test set coverage is limited, and models may exhibit dangerous behaviors outside known test scenarios that are not captured.57% similarUnverified该研究方法的核心局限是N1阶段的低灵敏度（low sensitivity）56% similarUnverifiedIn domain-specific NER tasks, human-machine aggregated agreement metrics (such as high kappa values) can be misleading, and the choice of evaluation metric must match task complexity.56% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1037863API
curl https://kongchang.com/api/v1/knowledge/claims/1037863MCP
get_claim(id=1037863)