Unverified50% confidenceFactExact time
Scale autoregression switches the unrolling dimension from time steps to temporal resolutions, with coarse-scale predictions serving as conditional inputs to guide fine-scale generation.
1
Sources
50%
Confidence
Long-term
Relevance
11/17/2026
First Seen
Sources
Related Entities
Related Claims
Unverified通过向前向传播激活中添加或减去特定方向向量,可以实时调节模型在该概念维度上的表现,无需修改权重参数65% similarUnverifiedThe basic idea of diachronic word vectors is to train separate vector spaces on corpora from different historical periods, then map them to the same coordinate system via alignment techniques for comparison.65% similarUnverifiedThe era of blindly stacking large-model parameters is, at least in the time-series forecasting direction, coming to an end.64% similarUnverifiedEmbedding models map raw content into a high-dimensional vector space, where the output vectors geometrically preserve semantic relationships, with content of similar meaning corresponding to vectors that are closer in distance.64% similarUnverifiedScaling laws reveal a near power-law relationship between model loss and the number of parameters, data volume, and compute63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1179841API
curl https://kongchang.com/api/v1/knowledge/claims/1179841MCP
get_claim(id=1179841)