Unverified50% confidenceFactExact time
Current mainstream large language models (such as GPT, Claude, etc.) adopt a session-based interaction mode; the context window has expanded from an initial 4K tokens to 128K or even longer, but it remains essentially a single-conversation memory mechanism, and implicitly accumulated knowledge is lost once a new session begins.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/7/2026
First Seen
Valid until: 12/6/2026
Sources
Related Entities
Related Claims
Verified大语言模型在长时间对话中会因上下文窗口限制而丢失早期信息74% similarUnverifiedIn large language models, one English word typically corresponds to 1–2 tokens, one Chinese character typically corresponds to 1.5–2.5 tokens, and GPT-4's single-conversation context window is approximately 128,000 tokens.74% similarUnverifiedThe context windows of current mainstream large language models have expanded to hundreds of thousands or even millions of tokens, but under ultra-long contexts the model's attention weights on earlier information decay significantly (the "lost in the middle" problem).72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/870470API
curl https://kongchang.com/api/v1/knowledge/claims/870470MCP
get_claim(id=870470)