Unverified50% confidenceFactExact time
Tiny-MoE实现了包括Top-k、Top-p、重复惩罚和n-gram屏蔽在内的文本生成策略
1
Sources
50%
Confidence
Long-term
Relevance
7/13/2026
First Seen
Sources
15岁少年从零训练200M MoE语言模型：完整架构与实战拆解
redditr/learnmachinelearning7/12/2026
Related Claims
Unverified项目实现了涵盖Top-k、Top-p、重复惩罚和n-gram屏蔽的文本生成策略，用于对抗小模型的重复循环缺陷77% similarUnverifiedLarge models process text using Token (word element) as the minimum processing unit, with tokenizers handling the segmentation.62% similarUnverifiedThe KV Cache mechanism can cache repeated prefix content, and the more concise the system prompt, the higher the cache hit rate.60% similarUnverified文本编码器将提示词编码为固定维度嵌入向量，abliteration针对的是token生成过程中的拒绝行为，与嵌入向量输出情况不同58% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/502163API
curl https://kongchang.com/api/v1/knowledge/claims/502163MCP
get_claim(id=502163)