Unverified50% confidenceFactExact time
Calculated in float16, a probability distribution over a vocabulary of 100K is approximately 200KB.
1
Sources
50%
Confidence
Long-term
Relevance
9/11/2026
First Seen
Sources
Related Entities
Related Claims
Unverified以英语约5万常用词汇计算，Trigram 理论上需存储约1.25×10¹⁴种组合的概率72% similarUnverifiedLLM生成阶段对词汇表（通常包含32,000至128,000个token）计算Softmax概率分布，形成总和为1的概率向量70% similarUnverifiedA model with 100 billion parameters requires approximately 80 GB of VRAM to process a single inference request, while a model with tens of billions of parameters requires only about 20 GB.69% similarUnverifiedGrok 1 的3140亿总参数中每次推理仅激活约25%，实际推理算力接近650亿参数规模的密集模型64% similarUnverifiedThe output of large models is probabilistic rather than deterministic; asking the same question 100 times may yield 100 different answers.64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/895485API
curl https://kongchang.com/api/v1/knowledge/claims/895485MCP
get_claim(id=895485)