Unverified50% confidenceFactExact time
The most common problem in on-device inference is memory, as context length silently inflates the KV cache until it collides with the weights.
1
Sources
50%
Confidence
Long-term
Relevance
10/12/2026
First Seen
Sources
Related Entities
Related Claims
VerifiedThe Ebbinghaus forgetting curve reveals that memory strength decays exponentially over time, and this process can be slowed through spaced repetition.77% similarUnverified在真实使用场景中反复遇到某个知识点，比集中死记硬背更能形成持久记忆，与间隔效应和必要难度理论相印证76% similarUnverified推理过程中动态增长的KV Cache会持续占用显存，长对话时可能突然触发OOM崩溃76% similarUnverifiedThe core contradiction in long-term memory design is: the more that is remembered, the higher the cost and noise, and the less timely the cleanup, the worse the accuracy — the industry currently has no standard answer.75% similarUnverifiedActive forgetting and summary compression are necessary measures to combat the unbounded growth of memory stores.75% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1005175API
curl https://kongchang.com/api/v1/knowledge/claims/1005175MCP
get_claim(id=1005175)