Unverified50% confidenceDecision RuleExact time
评测角色LoRA时应使用复杂提示词检验模型在全新场景中的泛化能力，而非仅测试记忆能力
1
Sources
50%
Confidence
Long-term
Relevance
7/17/2026
First Seen
Sources
Related Claims
UnverifiedBenchmarks tend to evaluate how much a model knows, rather than how much cognitive increment it can bring.75% similarUnverifiedSkillSpector supports two detection modes: local rule-based static scanning that requires no model, and deep semantic analysis that connects to a large language model.70% similarUnverifiedThe focus of evaluation is shifting from the quality of model text output to the ability of agents to complete end-to-end tasks, with evaluation dimensions expanding from a single accuracy metric to multi-dimensional metrics including task completion rate, step efficiency, error recovery capability, and time overhead.69% similarUnverifiedBenchmark hacking refers to models potentially being optimized for specific test sets rather than possessing general capabilities, resulting in a gap between benchmark scores and real-world performance.68% similarUnverified同一个模型用同一套假设检查自己的代码天然看不到盲点，跨模型审查因训练侧重点不同才具备互补价值68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/541285API
curl https://kongchang.com/api/v1/knowledge/claims/541285MCP
get_claim(id=541285)