How AI reads outing

1 Your word measured

“many summer excursions to the shore”

noun 5 senses 11 synonyms

Definitions come from WordNet, a hand-curated dictionary.

2 To an AI, it's pieces measured

DeepSeek V3 outing 1 token

Mistral 7B out ing 2 tokens

Qwen3 8B outing 1 token

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

scouting0.14 bully0.13 reunion0.13 enjoyable0.13 bout0.12 evenings0.12 excursion0.12 touring0.12

Dictionary synonyms measured

excursion0.12 picnic0.07 expedition0.06 out0.06 jaunt junket pleasure trip sashay field day come out of the closet come out

Shade shows how close the model puts each word to outing, 1 of 8 also appear in the dictionary. 4 of 11 dictionary synonyms are in this build; the rest have no vector to compare yet.

The model’s neighbours

excursion0.22 excursions0.19 outfit0.18 getaway0.18 trip0.17 outage0.16 listing0.16 outfits0.16

Dictionary synonyms measured

excursion0.22 expedition0.13 out0.12 picnic0.12 jaunt junket pleasure trip sashay field day come out of the closet come out

Shade shows how close the model puts each word to outing, 1 of 8 also appear in the dictionary. 4 of 11 dictionary synonyms are in this build; the rest have no vector to compare yet.

This model splits outing 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

out0.83 outlining0.76 outage0.67 outback0.64 outset0.62 outlying0.62 outward0.61 outlines0.60

Dictionary synonyms measured

out0.83 excursion0.14 picnic0.09 expedition0.05 jaunt junket pleasure trip sashay field day come out of the closet come out

Shade shows how close the model puts each word to outing, 1 of 8 also appear in the dictionary. 4 of 11 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 outing 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

rookie0.115 outsourced0.113 itinerary0.112 threesome0.112 sophomore0.111

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

outs0.159 outgoing0.158 trips0.158 outlet0.154 outset0.152

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

This model splits outing 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.

outlet0.600 outboard0.597 outgoing0.587 outlier0.575 outlets0.573

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

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