Unverified50% confidenceFactExact time
In traditional evaluations, models can rely on parametric knowledge or semantic cues in the compressed text to guess the answer even when the key evidence has been removed, a phenomenon called "shortcut learning."
1
Sources
50%
Confidence
Long-term
Relevance
10/20/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedCertain specific phrasings in the original prompt activate shortcuts entrenched during model training, causing the model to apply fixed patterns in its responses; after paraphrasing, the model is forced to perform more general reasoning, thereby improving output quality.76% similarUnverified知识图谱以工具（tools）形式暴露给前沿模型，而非直接塞进 prompt，使模型可按需多次调用查询76% similarUnverifiedIn semi-supervised learning in machine learning, unlabeled data usually requires additional assumptions or prior knowledge to be utilized.75% similarUnverifiedExplicitly naming Equibles in the prompt can prompt the model to call the tool rather than guess.75% similarVerifiedWhen input constraints are reduced, machine learning models rely more heavily on their prior distribution—that is, the default understanding of a concept formed during training.75% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1073298API
curl https://kongchang.com/api/v1/knowledge/claims/1073298MCP
get_claim(id=1073298)