Unverified50% confidenceFactExact time
The authors propose a new conditional entropy function as a regularization term that directly constrains the behavior of the evidence distribution, guiding the model to give sharper evidence distributions when the classification is correct and to retain sufficient ambiguity when uncertain.
1
Sources
50%
Confidence
Long-term
Relevance
11/11/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedIn a cascade architecture, the confidence condition is usually quantified by the probability distribution, entropy, or consistency with reference answers of the small model's output.73% similarVerifiedThe semantic entropy method is technically implemented by obtaining candidate answers through multiple samplings, using an NLI model to determine semantic equivalence for clustering, and then computing entropy over the semantic categories.73% similarUnverifiedThe semantic entropy method does not rely on the internal structure of a specific model; it infers uncertainty through semantic consistency at the output level, giving it relatively strong generalization ability.72% similarUnverifiedMost existing generative recommendation methods only use the standard causal cross-entropy objective to optimize semantic ID prediction, ignoring the hierarchical structure of the semantic IDs themselves.70% similarUnverifiedEffective branching factor is a concept in information theory and language modeling, referring to the number of nearly equally reasonable candidates at the next token position given the context, approximately characterized by the entropy of the distribution.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1158454API
curl https://kongchang.com/api/v1/knowledge/claims/1158454MCP
get_claim(id=1158454)