Verified75% confidenceFactExact time
Pass@1 represents the probability that the model generates a correct result in a single sampling, while Pass@k represents the probability of at least one correct result in k samplings.
3
Sources
75%
Confidence
Long-term
Relevance
9/18/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedHumanEval通过pass@k指标衡量模型生成正确函数实现的概率75% similarUnverifiedThe main theorem of the paper proves that for every fixed confidence level 1-δ, the risk is bounded by k/N + O(1/√N) with probability at least 1-δ, and this upper bound holds uniformly for all possible values of k.67% similarUnverifiedAdaptiveSpec's progressive marginal rule allows a draft token to be accepted when the ratio of the target model's probability on that draft token to its top-1 token probability exceeds a threshold, even if the two do not exactly match.64% similarVerifiedBecause the output of large models is probabilistic, the same input may produce different results across different runs.64% similarVerifiedLarge language models are fundamentally performing probabilistic predictions of the next most likely token in a sequence.64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/931852API
curl https://kongchang.com/api/v1/knowledge/claims/931852MCP
get_claim(id=931852)