Unverified50% confidenceFactExact time
Although step-by-step models train intermediate outputs and provide denser supervision signals, they still cannot fully offset the negative effects of module coverage shift.
1
Sources
50%
Confidence
Long-term
Relevance
11/11/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedOver-reliance on the fine-tuned part during fine-tuning may be precisely the root cause of rare-class capabilities being obscured; directing the model's computation moderately toward the pretrained components may be a more robust fine-tuning route.77% similarUnverified大模型对指令的敏感度呈非线性分布——主干指令的微小改动可能引发全局行为漂移,而在末尾追加具体的边界条件则相对安全74% similarUnverifiedIf the model memorized the answers to contaminated benchmarks during pre-training, information loss from compressing the context will not affect performance, rendering the reported 'lossless compression' conclusion meaningless.74% similarUnverifiedWeighted linear scoring combines multiple signals while remaining fast and interpretable, making it suitable for scenarios where the number of signals is manageable, training data is limited and interpretability matters, but its weakness is manual calibration.72% similarUnverified底层模型更换后,即便微小的输出分布变化也可能导致此前校准的安全护栏过度拦截或防护失效,需要重新标注和调试72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1158433API
curl https://kongchang.com/api/v1/knowledge/claims/1158433MCP
get_claim(id=1158433)