Unverified50% confidenceBenchmarkExact time
On the Banking77 dataset, CAPS achieves 2.6 percentage points higher accuracy than fine-tuning on confirmed answers only, 0.053 lower ECE, and lower selective risk (based on three random seeds versus one baseline seed).
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
11/13/2026
First Seen
Valid until: 2/11/2027
Sources
Related Entities
Related Claims
UnverifiedOn a set of ordered-scale questions, using probability scores as the REINFORCE reward caused accuracy to drop sharply from 0.636 to 0.111.69% similarUnverifiedAt N=5000, compared with full-label training, CAPS is 1.5 percentage points lower in accuracy on the 0.6B model and 0.6 percentage points lower on the 1.7B model (versus one full-label seed), and propensity weighting brought no gain on this log.68% similarUnverifiedOn the MedAbstain benchmark, the median appropriate abstention rate of the base models was only 8.3%, and after fine-tuning it further dropped to a median of 4.2%.68% similarUnverifiedThe Spearman correlation coefficient between tool description scores and correct model selection was only 0.04, with almost no statistical association.67% similarUnverifiedFor the 7B model, accuracy with the solver was 75.31% versus 72.02% for direct calculation, a paired improvement of +3.29 percentage points, with a 95% cluster interval of [-3.49, 10.38] crossing zero, making the advantage non-significant.67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1165000API
curl https://kongchang.com/api/v1/knowledge/claims/1165000MCP
get_claim(id=1165000)