Unverified50% confidenceFactExact time
Because the model operates on tokens rather than letters, tasks such as asking the model to count how many times a word appears or to reverse the letters of a word frequently result in errors.
1
Sources
50%
Confidence
Long-term
Relevance
10/2/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedEven models fine-tuned specifically for function calling can still produce parameter errors such as missing fields, numbers generated as strings, and out-of-range enum values.75% similarVerifiedBecause large language models are based on probabilistic sampling, even if instructions are written in the context, the model may still forget or skip certain instructions in complex reasoning chains, which is known as instruction-following instability.69% similarUnverified用户输入的每个词都会成为模型的上下文，改变模型接收到的数学输入，进而影响解码阶段的概率分布和输出内容69% similarUnverified对话式大模型输出的证明文字基于统计模式生成，无法保证每个步骤逻辑严格成立68% similarUnverifiedThe reasoning process of large models is billed based on the number of tokens consumed, with one token roughly equivalent to 3/4 of an English word68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/968542API
curl https://kongchang.com/api/v1/knowledge/claims/968542MCP
get_claim(id=968542)