Unverified50% confidenceFactExact time
When a language model computes a numerical average, running it twice in succession may yield different results (e.g., 49.6 and 49.8), both of which can pass schema validation but may not be correct.
1
Sources
50%
Confidence
Long-term
Relevance
10/11/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe sum of the negative log probabilities of each token output by a language model is an estimate of the encoding length of the sequence, which is the origin of the perplexity metric.69% similarVerifiedLarge model outputs are probabilistic and non-deterministic; running the same prompt on the same version of a model multiple times may produce different results, making quality judgments in gradual rollouts difficult.68% similarUnverified对同一测试用例执行多次调用统计成功比例得到通过率，比二元的通过/失败判断更能反映模型真实表现67% similarUnverified修复前模型每约29次下一token预测中约有1次的选择与数学上应有结果不同66% similarUnverified该研究发现同一模型不同部署时间点在代码生成和逻辑推理任务上通过率可能出现10%至20%的显著波动，且方向不总是单向改善66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/999965API
curl https://kongchang.com/api/v1/knowledge/claims/999965MCP
get_claim(id=999965)