Unverified50% confidenceFactExact time
推测解码由Google DeepMind的Leviathan等人和陈等人于2023年独立提出，基于拒绝采样的验证算法确保输出分布与主模型完全等价
1
Sources
50%
Confidence
Long-term
Relevance
7/20/2026
First Seen
Sources
Related Claims
VerifiedSpeculative decoding was formally proposed by Google DeepMind in 2023; overall throughput can be improved by 2–3 times, and the output quality is completely equivalent to that of the original large model.80% similarVerifiedThe speculative decoding paradigm was proposed almost simultaneously by teams at Google Brain and DeepMind in 2023.76% similarUnverified推测性解码由Google Brain于2023年正式提出，核心洞察是GPU在验证一批Token时的计算成本与验证单个Token相近76% similarUnverifiedSpeculative decoding was formally proposed by Google in 2023, using a draft model with 1/10 to 1/50 the parameters of the target model to generate multiple candidate tokens at once, which are then verified in parallel by the large model.75% similarUnverifiedGoogle DeepMind在2023年提出了能力弹性（Capability Elicitation）框架，强调模型的真实行为边界只有在特定触发条件下才会暴露72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/575549API
curl https://kongchang.com/api/v1/knowledge/claims/575549MCP
get_claim(id=575549)