Unverified50% confidenceFactExact time
Infinite-Parameter LLMs refer to a paradigm in which model weights are no longer static fixed values stored in memory, but can be dynamically generated from real-time data and continuously adjusted
1
Sources
50%
Confidence
Long-term
Relevance
9/17/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe autoregressive decoding phase of LLMs is limited by memory bandwidth bottlenecks, as generating each token requires a complete traversal of the activated model weights.69% similarUnverified随着模型参数规模从亿级扩展到千亿级，LLM展现出涌现能力，包括多步推理、类比迁移和不确定性表达68% similarUnverified大语言模型（LLM）在架构层面是无状态系统，每次推理调用是一次独立的前向传播计算，模型权重在推理阶段完全冻结67% similarUnverifiedLLM在推理阶段参数是冻结的，模型不会因用户交互而更新权重，因此本质上是无状态的67% similarUnverifiedLLM自回归解码阶段每生成一个token都需从显存加载全量模型权重，批量较小时形成内存墙67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/929997API
curl https://kongchang.com/api/v1/knowledge/claims/929997MCP
get_claim(id=929997)