Unverified50% confidenceFactExact time
In the Little Kiln Support Copilot tests, relevant historical cases scored above 0.63 in cosine similarity and noise cases below 0.58, so the author raised the retrieval threshold from 0.5 to 0.6.
1
Sources
50%
Confidence
Long-term
Relevance
10/18/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe experimental results show that the paired token cosine similarity is only 0.221–0.484.62% similarUnverifiedThe McNemar test is applicable to paired binary data, and the p=0.00006 (15 fixes, 0 breaks across the three packages combined) and p=0.03 (6 fixes, 0 breaks on the check packages only) reported in this paper both come from that test.61% similarUnverifiedOn real storage, the top score is 0.635 and the tail is only 0.450; in the 0.185 gap only the first row deserves attention, and everything below 0.55 is a noise band.59% similarUnverifiedTencent's published comparison table shows Hy-Memory-30B-A3B scoring an average of 55.49 across six benchmarks, slightly above Claude Opus 5's 54.62, a gap of 0.87 points; but looking at individual items, 0.02 points under the Hy-Memory framework and 0.01 points under MemOS are essentially noise level.59% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1063817API
curl https://kongchang.com/api/v1/knowledge/claims/1063817MCP
get_claim(id=1063817)