Feed Entities Podcasts Deep Dives Developers About

[KongchangAI]

Real-time AI knowledge base. Every claim is traceable to its source, with data freshness under 24 hours.

Powered by 240+ first-hand AI sources

Knowledge

Feed
Hot Entities
Podcasts

Developers

CLI
MCP Server
REST API
About

© 2026 KongchangAI. The freshest, most accurate AI knowledge

Unverified50% confidenceFactExact time

微学习（Microlearning）模式源自认知负荷理论，将复杂信息拆解为5-10分钟的小单元能提升知识保留率

1

Sources

50%

Confidence

Long-term

Relevance

8/19/2026

First Seen

Sources

AdmitRaven：大学申请界的Duolingo，每日微课替代万元顾问

Related Claims

Unverified推理能力可通过知识蒸馏和强化学习微调压缩，但事实性知识的存储依赖参数规模，压缩会导致知识覆盖面下降和幻觉率上升75% similar Unverified在4GB显存上微调伴随训练速度极慢和微调能力受限的代价，更适合小规模数据的轻量适配而非大规模知识注入73% similar Unverified丰富的上下文信息能收窄模型输出概率分布并减少幻觉，这一现象被称为「知识锚定效应」72% similar Unverified经过优化的紧凑提示词往往比冗长提示词效果更好，因为冗余信息会分散模型的注意力权重69% similar Unverified在推理阶段内存带宽往往比纯算力更早成为瓶颈，注意力机制需在每步生成时读取全部KV缓存，上下文长度增大时内存读写开销呈平方级增长68% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/774526

API

curl https://kongchang.com/api/v1/knowledge/claims/774526

MCP

get_claim(id=774526)