Unverified50% confidenceFactExact time
研究表明人类阅读字幕时单次注视有效信息量约为7±2个汉字，超过14-16字单行字幕会显著增加认知负荷
1
Sources
50%
Confidence
Long-term
Relevance
7/7/2026
First Seen
Sources
AI Agent Fully Automates TV Subtitle Generation: From Audio Extraction to Final Output in One Go
bilibililonrenyt7/5/2026
Related Claims
Unverified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70% similarUnverifiedLarge language models are billed by the number of tokens, with 1 token equaling approximately 4 English characters or 1–2 Chinese characters.69% similarUnverified根据 Simon Willison 实测，新分词器导致英文文本 Token 增幅 +42%，西班牙文 +33%，Python 代码 +27%，中文基本无变化68% similarUnverifiedThe context window of large language models is limited, and the length of text that can be processed in a single inference is typically tens of thousands to hundreds of thousands of tokens.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/167917API
curl https://kongchang.com/api/v1/knowledge/claims/167917MCP
get_claim(id=167917)