Unverified50% confidenceFactExact time
In a binary classification task, if the positive-to-negative sample ratio is 6:4, then a baseline that always outputs "positive" scores 60 points, and an agent scoring 65 has only 5 points of real signal.
1
Sources
50%
Confidence
Long-term
Relevance
10/15/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedUnder the condition of n=100 with an even split of positive and negative samples, the 95% confidence interval for 55% accuracy is roughly 45%–65%, which covers the 50% baseline, so the 5 percentage point gap may be random fluctuation.70% similarUnverifiedIn the specific experimental case, the overlapping-anchor scheme (n=912) showed the model at 55.2% versus the majority class at 51.1%, a net advantage of +4.1pp, about 2.4 standard errors; the non-overlapping-anchor scheme (n=163) showed 59.5% versus 54.0%, a net advantage of +5.5pp, about 1.4 standard errors.70% similarUnverifiedThe benchmark produces two numbers for each model that are never merged: a task score and a false-confidence rate (the proportion of cases that should have been escalated but were answered anyway).69% similarUnverifiedRanked by question difficulty, persistence rate was 91.4% on the 60 questions that all models answered correctly, but only 73.5% on the 52 questions that at least one model got wrong, indicating that sycophancy tracks uncertainty.69% similarUnverifiedFairness instructions had limited effect: family accuracy changed by only −0.3 percentage points on the cluster six-model panel and +2.5 percentage points on Kaggle, with confidence intervals spanning zero in both cases.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1037052API
curl https://kongchang.com/api/v1/knowledge/claims/1037052MCP
get_claim(id=1037052)