1
Your word
measured
“the probability that an unbiased coin will fall with the head up is 0.5”
Definitions come from WordNet, a hand-curated dictionary.
a measure of how likely it is that some event will occur; a number expressing the ratio of favorable cases to the whole number of cases possible
几率 is Chinese for probability, and every measurement below was made on that English word. Definitions come from English WordNet; Open Multilingual WordNet supplies the Chinese headword, not a Chinese definition.
Panels 3, 4 and 5 are computed per model. Switch to see them disagree.
“the probability that an unbiased coin will fall with the head up is 0.5”
Definitions come from WordNet, a hand-curated dictionary.
2 of those are a raw byte rather than a character: Qwen3 8B splits this word below the character, so no single token it sees is readable.
This panel splits 几率 itself, not probability — a tokenizer does not care what language it is fed. It is the only panel here that does.
This is the tokenizer splitting text, not the model understanding it.
Shade shows how close the model puts each word to probability, 1 of 8 also appear in the dictionary.
Shade shows how close the model puts each word to probability, 1 of 8 also appear in the dictionary.
Shade shows how close the model puts each word to probability.
Model neighbours are distributional, not dictionary synonyms. Two words can be close because they appear in similar sentences, which is why an antonym can outscore a synonym here.
Scores are projections onto axes we defined from anchor words, not labels the model assigns.
These sit just outside the closest neighbours in panel 3, and no dictionary lists any of them as related to probability. That gap is the model’s own learned association.
These sit just outside the closest neighbours in panel 3, and no dictionary lists any of them as related to probability. That gap is the model’s own learned association.
These sit just outside the closest neighbours in panel 3, and no dictionary lists any of them as related to probability. That gap is the model’s own learned association.
These are statistical associations in the training data, not the model thinking.