Unverified50% confidenceFactExact time
Anari et al. (2026) proposed using rejection sampling to correct the distributional bias in parallel decoding of masked diffusion models, decomposing the log acceptance probability into a sum of token-level terms.
1
Sources
50%
Confidence
Long-term
Relevance
10/16/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe verification process in traditional speculative decoding accepts draft-proposed tokens with probability min(1, p(x)/q(x)), and upon rejection resamples from the corrected distribution norm(max(0, p(x)-q(x))), which can be proven to produce a token distribution identical to directly sampling from the target model (lossless).68% similarUnverifiedQi et al. (ICLR 2024) observed that as few as a small number of adversarial samples can cause model alignment degradation, with even benign datasets producing a smaller negative effect.63% similarUnverifiedWhen the prediction error is assumed to follow a Gaussian distribution with zero mean and fixed variance, taking the negative logarithm of the likelihood function and expanding it yields a leading term that is exactly in the form of mean squared error.63% similarUnverified在某些测试条件下，主流AI编程工具推荐不存在包名的概率可达5%~20%62% similarUnverifiedExperiments show that models processed with abliteration exhibit almost no degradation on standard capability evaluations, but the probability of refusing harmful requests drops significantly.62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1042093API
curl https://kongchang.com/api/v1/knowledge/claims/1042093MCP
get_claim(id=1042093)