Unverified50% confidenceFactExact time
After each round of training, AdaBoost increases the weights of misclassified samples and decreases the weights of correctly classified samples.
1
Sources
50%
Confidence
Long-term
Relevance
10/5/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe quality ceiling of weak supervision labels depends on the accuracy of the zero-shot model itself; if the model makes systematic errors, the training set will inherit that bias.73% similarUnverifiedInfoNCE 损失通过在一批候选中最大化正样本对相似度、最小化负样本对相似度进行训练，温度超参数τ越小分布越集中70% similarUnverifiedVendor-reported coding scores are typically measured on carefully curated subsets, and results reproduced by independent organizations on the full test set are often lower.68% similarUnverifiedArtifacts produced through learning can reduce the number of samples required to reach a given score at test time.68% similarUnverifiedAnthropic's research shows that even mild alignment training can cause a 2–8 percentage point decline in model performance on certain reasoning-intensive tasks.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/977406API
curl https://kongchang.com/api/v1/knowledge/claims/977406MCP
get_claim(id=977406)