Unverified50% confidenceTradeoffExact time
Lexical overlap metrics are fast to compute, reproducible, and deterministic, but may underestimate semantically correct answers that involve synonym substitution, word-order changes, or paraphrasing.
1
Sources
50%
Confidence
Long-term
Relevance
9/16/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedA lower --stylize value makes the generated result more faithful to the literal description of the prompt, while a higher value causes the model to introduce more aesthetic judgment but deviate from the precise instructions of the prompt.71% similarUnverifiedMost LLMs exhibit a 'verbosity preference,' tending to rate longer and more detailed answers higher even if those answers are not necessarily more accurate71% similarUnverifiedIn few-shot prompting, simply changing the order of examples can cause dramatic fluctuations in accuracy, ranging from near-random to near-optimal.70% similarUnverifiedIn machine learning evaluation, precision and recall often have a trade-off relationship, and increasing recall typically leads to more genuine products being misclassified.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/927975API
curl https://kongchang.com/api/v1/knowledge/claims/927975MCP
get_claim(id=927975)