Unverified50% confidenceBenchmarkExact time
In the Pyrite benchmark evaluation, PyriteReasoner and PyriteFeaturizer achieved ranking accuracies of 67% and 74%, respectively.
1
Sources
50%
Confidence
Long-term
Relevance
11/11/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe complete system achieved an accuracy of 77.3% across six simulator-derived report fields.74% similarUnverifiedJoint adaptation raises the four-benchmark macro-average accuracy from 66.46% to 74.32%.68% similarUnverifiedOn the 531 checked answers, Scribe's final configuration achieved 96% of answers read completely correctly, up from 93%.67% similarUnverifiedIn schemagate's benchmark, on the Spider dataset's 1034 questions the pure mathematical ranking accuracy is 80.3%, and 96.5% after re-ranking; on the BIRD dataset's 1534 questions the pure mathematical ranking is 88.6%, and 96.7% after re-ranking.66% similarUnverifiedCompared with a strong adaptive allocation baseline, this method achieved absolute accuracy improvements of up to 3.73 percentage points (LLaDA) and 7.33 percentage points (Dream) on MATH-500.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1159155API
curl https://kongchang.com/api/v1/knowledge/claims/1159155MCP
get_claim(id=1159155)