Unverified50% confidenceSolutionExact time
Model output should be treated entirely as untrusted user input and subjected to encoding, parameterization, and validation.
1
Sources
50%
Confidence
Long-term
Relevance
10/11/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe user stated that the models require extremely detailed prompts, and even when a standardized prompt is provided, the output remains difficult to trust.74% similarUnverifiedSpeculative decoding improves generation throughput while guaranteeing that the output is completely identical to that of the original model.72% similarUnverifiedThere are two paths to getting the model to output an accurate solution: writing detailed requirements all at once, or using follow-up questioning to have the model think on the user's behalf and accumulate context.71% similarUnverifiedThe description field in a tool definition must be written clearly, because it is precisely what the model relies on to determine when to call the function.71% similarUnverified应将大模型输出视为外部输入，必须经过完整校验才能进入核心业务，而非当作可信数据71% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/998097API
curl https://kongchang.com/api/v1/knowledge/claims/998097MCP
get_claim(id=998097)