Unverified50% confidenceSolutionExact time
Because each model may use a different tokenizer, it is not possible to directly pass the model's hidden states; the mainstream solution is to use natural language as an intermediate representation for context transfer.
1
Sources
50%
Confidence
Long-term
Relevance
9/5/2026
First Seen
Sources
Related Entities
Related Claims
Unverified模型的工具调用能力来自专门的微调训练，而非通用语言模型的自然涌现71% similarVerifiedLarge language models are based on the Transformer architecture; each inference is an independent forward propagation process that retains no session state.68% similarUnverified大语言模型采用自回归生成机制，在生成token时使用单向前向注意力，无法进行全局类型校验，易产生API幻觉68% similarUnverifiedThe tool-calling mechanism allows large language models to dynamically invoke external functions during inference, representing a fundamental leap from language models to action models.68% similarUnverified模糊的感受类描述会让大语言模型在多个可能的方向上随机游走，而具体的问题描述则直接指向解决路径68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/859152API
curl https://kongchang.com/api/v1/knowledge/claims/859152MCP
get_claim(id=859152)