How a language model learns: guessing the next word, sentence 0
A chatbot is built on a "large language model": large because it is enormous, language because it works with words, model because it is a copy of the patterns in how people write.
Who acts: chatbot, people · What happens: build, be, work, write · To what, where: chatbot, large enormous language, copy
This is a sentence of the lesson «How a language model learns: guessing the next word» of the ai course of Mova's primary school. The page gives its syntax tree — a dependency tree in the style of Universal Dependencies — and its pragmatics clause by clause: who acts (chatbot, people), what happens, when and where, as Mova's frames read it from the tree.
The tree
The words, one row each (CoNLL-U)
| # | word | lemma | UPOS | features | head | relation |
|---|---|---|---|---|---|---|
| 1 | A | a | DET | Definite=Ind PronType=Art | 2 chatbot | det |
| 2 | chatbot | chatbot | NOUN | Number=Sing | 4 built | nsubj:pass |
| 3 | is | be | AUX | Mood=Ind Number=Sing Person=3 Tense=Pres VerbForm=Fin | 4 built | aux:pass |
| 4 | built | build | VERB | Tense=Past VerbForm=Part Voice=Pass | root | root |
| 5 | on | on | ADP | _ | 10 model | case |
| 6 | a | a | DET | Definite=Ind PronType=Art | 10 model | det |
| 7 | " | " | PUNCT | _ | 10 model | punct |
| 8 | large | large | ADJ | Degree=Pos | 10 model | amod |
| 9 | language | language | NOUN | Number=Sing | 10 model | compound |
| 10 | model | model | NOUN | Number=Sing | 4 built | obl |
| 11 | " | " | PUNCT | _ | 10 model | punct |
| 12 | : | : | PUNCT | _ | 19 language | punct |
| 13 | large | large | ADJ | Degree=Pos | 19 language | amod |
| 14 | because | because | SCONJ | _ | 17 enormous | mark |
| 15 | it | it | PRON | Case=Nom Gender=Neut Number=Sing Person=3 PronType=Prs | 17 enormous | nsubj |
| 16 | is | be | AUX | Mood=Ind Number=Sing Person=3 Tense=Pres VerbForm=Fin | 17 enormous | cop |
| 17 | enormous | enormous | ADJ | Degree=Pos | 19 language | amod |
| 18 | , | , | PUNCT | _ | 19 language | punct |
| 19 | language | language | NOUN | Number=Sing | 10 model | appos |
| 20 | because | because | SCONJ | _ | 22 works | mark |
| 21 | it | it | PRON | Case=Nom Gender=Neut Number=Sing Person=3 PronType=Prs | 22 works | nsubj |
| 22 | works | work | VERB | Mood=Ind Number=Sing Person=3 Tense=Pres VerbForm=Fin | 19 language | acl:relcl |
| 23 | with | with | ADP | _ | 24 words | case |
| 24 | words | word | NOUN | Number=Plur | 22 works | obl |
| 25 | , | , | PUNCT | _ | 26 model | punct |
| 26 | model | model | NOUN | Number=Sing | 24 words | appos |
| 27 | because | because | SCONJ | _ | 31 copy | mark |
| 28 | it | it | PRON | Case=Nom Gender=Neut Number=Sing Person=3 PronType=Prs | 31 copy | nsubj |
| 29 | is | be | AUX | Mood=Ind Number=Sing Person=3 Tense=Pres VerbForm=Fin | 31 copy | cop |
| 30 | a | a | DET | Definite=Ind PronType=Art | 31 copy | det |
| 31 | copy | copy | NOUN | Number=Sing | 4 built | ccomp |
| 32 | of | of | ADP | _ | 34 patterns | case |
| 33 | the | the | DET | Definite=Def PronType=Art | 34 patterns | det |
| 34 | patterns | pattern | NOUN | Number=Plur | 31 copy | nmod |
| 35 | in | in | SCONJ | _ | 38 write | mark |
| 36 | how | how | ADV | PronType=Int | 38 write | advmod |
| 37 | people | people | NOUN | Number=Plur | 38 write | nsubj |
| 38 | write | write | VERB | Mood=Ind Number=Plur Person=3 Tense=Pres VerbForm=Fin | 4 built | advcl |
| 39 | . | . | PUNCT | _ | 4 built | punct |
Pragmatics
- form
- statement
- act
- assert
- voice
- narrator
- polarity
- none
- teacher
- changed act=assert
| # | verb | tense, mood | who | what, whom | to whom | where | when | other | roles by |
|---|---|---|---|---|---|---|---|---|---|
| 1 | build | present · passive | a chatbot | on a large language model | vn base-97.1 «NP V NP PP.source»: subject=Theme on model=Source | ||||
| 2 | is | present | chatbot | enormous property | tree rules | ||||
| 3 | work | present | chatbot | large enormous language not said here | with words | vn knead-26.5 «NP V NP.material»: agent=Agent language=Material | |||
| 4 | is a | present | chatbot | a copy | of the patterns | tree rules | |||
| 5 | write | present | people | vn scribble-25.2-1 «NP V»: agent=Agent |
One row per clause the reader finds. «not said here»: a role filled from the clause above. «roles by»: a VerbNet class and frame when one fits the verb, else the reader's rules on the tree.
Licence: CC BY-SA 4.0, as the whole of Mova corpus 0.4, of which this treebank (the agreed school, mova-school-agreed-en.conllu) is part. The lessons were written for Mova's primary school. Sentence school:ai:002-guessing-the-next-word.lesson:0.
The tree: Mova annot: Rust (en) → Haiku 5.5 (medium, Batches API) patches; gates UD 2.18; agreed by treeagree rules (Stanza, Trankit, spaCy trf): R6-punct-agree ×1. The frames are read from the tree by Mova's frames reader (prag) when this page is built, the lesson read in order, so a pronoun may be resolved from an earlier sentence.
See the release 0.4.
This page in Markdown → https://md.mova.elamur.ai/free/en/corpora/school/ai/2-guessing-the-next-word/0/