Unverified50% confidenceFactExact time
The training objective of quantile regression is the Pinball Loss; for target quantile τ, predictions above the true value are penalized with weight (1-τ), and predictions below are penalized with weight τ.
1
Sources
50%
Confidence
Long-term
Relevance
10/5/2026
First Seen
Sources
Related Entities
Related Claims
Unverified分配给预训练的算力越多，下游任务的pass@1奖励就越高，预训练损失越低模型在后续RL中获得的收益越大66% similarUnverifiedDuring the post-training phase, an excessively low loss may correspond to overfitting, where the model memorizes the training data and loses generalization ability.66% similarUnverifiedThe student model's training objective is a weighted combination of a conventional loss against true hard labels and a distillation loss against the teacher's soft labels, with the distillation loss generally measured by KL divergence.65% similarUnverifiedThe paper finds that the ranking of downstream tasks is not consistent with the ranking of training loss, and schemes with lower training loss do not necessarily perform better in downstream evaluations.64% similarUnverifiedResearch such as Microsoft's BitNet shows that if 1-bit quantization-aware training (QAT) is applied from the pre-training stage, accuracy loss can be controlled, whereas post-training quantization (PTQ) typically incurs greater loss.62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/975952API
curl https://kongchang.com/api/v1/knowledge/claims/975952MCP
get_claim(id=975952)