Unverified50% confidenceFactExact time
DFQI's L1 penalty induces sparsity, helping identify which source information is most valuable for the target.
1
Sources
50%
Confidence
Long-term
Relevance
11/3/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedLLM幻觉问题源于模型基于概率预测生成文本而非检索真实数据，在涉及具体统计数字、引文来源和法律条款时尤为危险64% similarUnverifiedCaching is most useful at the tool-call layer but dangerous at the LLM inference layer—if only the prompt template is used as the cache key while ignoring the current state, stale inference results will be returned.64% similarUnverifiedHigh accuracy in LLM evaluations may be a product of information leakage rather than genuine reasoning ability, a problem that is especially pronounced in structured scientific data such as CIF files, which simultaneously contain multiple attribute fields.62% similarUnverified检索策略的效果高度依赖数据特征——在FAQ上BM25可能打败向量检索，在语义模糊的查询上则相反。62% similarUnverifiedKey metrics for measuring real-time threat intelligence sources include timeliness, coverage, confidence labeling, and machine-readability.62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1129563API
curl https://kongchang.com/api/v1/knowledge/claims/1129563MCP
get_claim(id=1129563)