Unverified50% confidenceBenchmarkExact time
On a misleading advertisement detection task, SJR achieved a 23.6% relative improvement in F1 score on the 'non-misleading' category compared to the zero-shot chain-of-thought baseline.
1
Sources
50%
Confidence
Long-term
Relevance
9/22/2026
First Seen
Sources
Related Entities
Related Claims
Unverified在评估阶段对候选关系标签进行随机打乱后，该模型的F1从37%骤降到14%70% similarUnverifiedOne SJR training variant uses no real violating samples at all (all positive-class data is synthetically generated), and its violation-class F1 score differs from the model trained on full real data by less than 0.2% in relative terms.67% similarUnverifiedEven the best-performing model achieves an evidence ID F1 of only around 0.36, reflecting that current large models still have significant room for improvement on cross-disciplinary reasoning tasks requiring precise evidence localization.62% similarUnverifiedAnthropic's Auto Mode classifier had a false positive rate of only 0.4% in testing on ten thousand real internal traffic samples, incorrectly blocking only 40 items.62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/940017API
curl https://kongchang.com/api/v1/knowledge/claims/940017MCP
get_claim(id=940017)