Unverified50% confidenceFactExact time
Neither Recall@k nor Reciprocal Rank@k measures whether the final answer is correct, safe, or grounded; both only answer questions at the retrieval level.
1
Sources
50%
Confidence
Long-term
Relevance
10/12/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe retrieval layer and generation layer should be evaluated separately; otherwise, a single answer quality score cannot indicate where to investigate.72% similarUnverifiedThe pragmatic approach is to first evaluate retrieval quality independently before connecting the generation stage: if the Top-K retrieved content does not contain the answer, switching to a stronger model cannot remedy the situation.69% similarUnverifiedBy appropriately choosing τ and c, one can selectively recover either the mean or the variance of the NNGP posterior, but cannot simultaneously recover all statistical properties of the exact posterior without trade-offs.65% similarUnverifiedA local question-answering mechanism based on paragraph retrieval and matching carries no hallucination risk and answers come directly from the source text, but it cannot perform cross-paragraph reasoning or synthesize and summarize information.64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1005604API
curl https://kongchang.com/api/v1/knowledge/claims/1005604MCP
get_claim(id=1005604)