Verified80% confidenceFactTime unknown
The vocabulary of a standard autoregressive language model typically contains 32,000 to 128,000 tokens.
4
Sources
80%
Confidence
Long-term
Relevance
6/1/2026
First Seen
Sources
Related Claims
Unverified现代语言模型的词汇表规模通常在32,000到100,000个token之间，使用BPE或SentencePiece等分词方法78% similarUnverifiedThe English version of Layla has 420 million parameters, the multilingual version has 320 million parameters and covers hundreds of languages, and there is an additional version fine-tuned for workflows.68% similarUnverified最新的大语言模型支持12.8万到100万token的上下文，但处理数十万字长篇小说时仍会出现信息遗忘和前后矛盾67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/21216API
curl https://kongchang.com/api/v1/knowledge/claims/21216MCP
get_claim(id=21216)