baño

noun 5 senses

a room or building equipped with one or more toilets

baño is Spanish for john, and every measurement below was made on that English word. Definitions come from English WordNet; Open Multilingual WordNet supplies the Spanish headword, not a Spanish definition.

also covers bathroom toilet tub bathrooms coating coatings
Story Show the data
English Español cabinet お手洗い 中文
Qwen3 8B
DeepSeek V3
Mistral 7B

Panels 3, 4 and 5 are computed per model. Switch to see them disagree.

1 Your word measured
noun 5 senses 18 synonyms

Definitions come from WordNet, a hand-curated dictionary.

2 To an AI, it's pieces measured
Qwen3 8B baño 1 token
DeepSeek V3 ba ño 2 tokens
Mistral 7B ba ño 2 tokens

This panel splits baño itself, not john — 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.

3 Synonyms, antonyms, and the AI measured
Qwen3 8B
DeepSeek V3
Mistral 7B

The model’s neighbours

johnston0.71 johns0.71 johnson0.69 johnny0.67 robert0.17 robertson0.15 james0.15 david0.14

Dictionary synonyms measured

bathroom0.06 trick0.05 toilet0.04 can0.03 lavatory lav privy King John John Lackland Saint John St. John Saint John the Apostle St. John the Apostle John the Evangelist John the Divine whoremaster whoremonger Gospel According to John

Shade shows how close the model puts each word to john. 4 of 18 dictionary synonyms are in this build; the rest have no vector to compare yet.

The model’s neighbours

johnson0.78 johnny0.77 johns0.75 johnston0.73 jonah0.19 jonathan0.19 jack0.19 jackson0.18

Dictionary synonyms measured

toilet0.07 bathroom0.06 can0.03 trick0.01 lavatory lav privy King John John Lackland Saint John St. John Saint John the Apostle St. John the Apostle John the Evangelist John the Divine whoremaster whoremonger Gospel According to John

Shade shows how close the model puts each word to john. 4 of 18 dictionary synonyms are in this build; the rest have no vector to compare yet.

This model splits john into 2 pieces, so it has no vector of its own here: this is the average of its fragments, and the results below are correspondingly rough.

The model’s neighbours

johns0.94 johnson0.88 johnny0.86 johnston0.84 jones0.56 jive0.56 jay0.55 jets0.55

Dictionary synonyms measured

toilet0.03 bathroom0.02 trick0.02 can0.02 lavatory lav privy King John John Lackland Saint John St. John Saint John the Apostle St. John the Apostle John the Evangelist John the Divine whoremaster whoremonger Gospel According to John

Shade shows how close the model puts each word to john. 4 of 18 dictionary synonyms are in this build; the rest have no vector to compare yet.

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.

4 Its personality measured
Qwen3 8B
DeepSeek V3
Mistral 7B
rare common
concrete abstract
casual formal
everyday technical
negative positive
mild intense
powerless powerful
small big
rare common
concrete abstract
casual formal
everyday technical
negative positive
mild intense
powerless powerful
small big

This model splits john into 2 pieces, so it has no vector of its own here: this is the average of its fragments, and the results below are correspondingly rough.

rare common
concrete abstract
casual formal
everyday technical
negative positive
mild intense
powerless powerful
small big

Scores are projections onto axes we defined from anchor words, not labels the model assigns.

5 Surprising neighbours measured
Qwen3 8B
DeepSeek V3
Mistral 7B
jan0.144 michael0.137 january0.134 roberts0.131 china0.123

These sit just outside the closest neighbours in panel 3, and no dictionary lists any of them as related to john. That gap is the model’s own learned association.

jan0.170 christian0.160 blackjack0.157 american0.150 january0.150

These sit just outside the closest neighbours in panel 3, and no dictionary lists any of them as related to john. That gap is the model’s own learned association.

This model splits john into 2 pieces, so it has no vector of its own here: this is the average of its fragments, and the results below are correspondingly rough.

james0.551 jock0.550 jig0.549 jade0.544 jong0.544

These sit just outside the closest neighbours in panel 3, and no dictionary lists any of them as related to john. That gap is the model’s own learned association.

These are statistical associations in the training data, not the model thinking.