Unverified50% confidenceOpinionExact time
Vector embeddings measure "how similar" rather than "whether connected," and therefore provide no guarantee for determining calls and dependencies.
1
Sources
50%
Confidence
Long-term
Relevance
10/15/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedEvaluation metrics based on semantic similarity (such as cosine similarity) cannot capture differences at the level of behavioral constraints and have systematic blind spots.72% similarUnverifiedUncensoring primarily affects a model's tendency to refuse, rather than its fundamental reasoning and generation capabilities, so capability loss is not significant for alignment-insensitive tasks such as coding.71% similarUnverifiedAblation experiments show that relying solely on reasoning yields no significant improvement in accuracy for rare entities.71% similarUnverifiedA benchmark can yield a perfect score while not measuring at all the difficult behavior you care about.71% similarUnverifiedAny analysis that attempts to explain human behavior using purely quantitative metrics may reach biased conclusions by ignoring utility dimensions that are difficult to measure.71% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1040811API
curl https://kongchang.com/api/v1/knowledge/claims/1040811MCP
get_claim(id=1040811)