Unverified50% confidenceFactExact time
Laya is a non-autoregressive decision model with 421M parameters.
1
Sources
50%
Confidence
Long-term
Relevance
9/18/2026
First Seen
Sources
Related Entities
Related Claims
VerifiedDense架构中模型所有参数在每次推理时都参与计算，计算成本随参数规模线性增长，以Llama系列为代表58% similarUnverifiedAlthough LLMs do not directly output normalized probability vectors, decision confidence can be approximately estimated through various engineering methods.57% similarUnverifiedLaya is an open-source decision model evolved from the Jev architecture that surpasses Jev on multiple Jev benchmarks through the RLCD training method.57% similarUnverifiedL2 regularization (Ridge regression) is equivalent, from a Bayesian perspective, to assuming that model parameters follow a Gaussian prior distribution with a mean of zero.56% similarUnverifiedSpeculative decoding is a lossless acceleration technique that does not alter the model's output distribution; it uses a draft model with far fewer parameters than the target model to generate candidate tokens, which the target model then verifies in parallel.56% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/931930API
curl https://kongchang.com/api/v1/knowledge/claims/931930MCP
get_claim(id=931930)