Unverified50% confidenceBenchmarkExact time
CacheGraft outperformed the plain-text receiver baseline in 29 of 30 comparisons, of which 17 were statistically significant.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
11/6/2026
First Seen
Valid until: 2/4/2027
Sources
Related Entities
Related Claims
UnverifiedAblation experiments show that, at the same precision, soft-label hierarchical retention keeps 751% more candidate words than strict power gating, expanding the candidate pool from 41 words to 349 words.70% similarUnverifiedIn tw-exam-bench, the answer-key consistency rate for 31,312 groups of duplicated questions reached 99.91%, and the only 27 discrepancies were verified and attributed one by one.70% similarUnverifiedAmong published results on SemBench, 21% of text results vary in quality across different executions, including engines whose query plans are themselves deterministic65% similarUnverifiedThe paper evaluates TextSeal2 on three LLMs; at a false positive rate of 10^-3 and equal output token diversity, the detection rate for 100-token answers is 8 to 22 percentage points higher than TextSeal.65% similarUnverifiedOn invalid propositions, using verifier feedback yielded 23/39 accepted counterexamples, while generic feedback without a checker yielded only 15/39.64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1141595API
curl https://kongchang.com/api/v1/knowledge/claims/1141595MCP
get_claim(id=1141595)