Unverified50% confidenceFactExact time
Joint accuracy refers to the proportion of responses where the model's answer is both correct and given in the target language, penalizing both language errors and answer errors at once, and thus more faithfully reflecting multilingual reasoning ability.
1
Sources
50%
Confidence
Long-term
Relevance
11/4/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedTraditional large language model evaluation centers on the accuracy of single-turn question answering, scoring the output answer against a reference answer after a question is given77% similarVerifiedLarge language models can approximately estimate decision confidence by analyzing token probability distributions, explicitly outputting self-evaluation scores, or observing answer consistency across multiple samplings.75% similarUnverifiedThe robustness of answer parsing often directly determines the evaluation score, because a model may express the same answer in multiple formats.71% similarUnverifiedCalibration measures the degree of consistency between the confidence a model expresses and its actual accuracy.70% similarUnverifiedThe accuracy of local models is generally proportional to model size; larger models require more computing resources, and users need to balance speed and accuracy based on their hardware conditions.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1134681API
curl https://kongchang.com/api/v1/knowledge/claims/1134681MCP
get_claim(id=1134681)