How a language model learns: guessing the next word, sentence 2

Now, when you ask something, it builds an answer a little piece at a time, each time picking a word that fits well after the words before it.

Who acts: you, chatbot, word · What happens: ask, build, pick, fit · To what, where: something, answer, word, chatbot

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 (you, chatbot, word), 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	Now	now	ADV	_	9 builds	advmod
2	,	,	PUNCT	_	1 Now	punct
3	when	when	ADV	_	5 ask	advmod
4	you	you	PRON	Case=Nom Person=2 PronType=Prs	5 ask	nsubj
5	ask	ask	VERB	Mood=Ind Number=Sing Person=2 Tense=Pres VerbForm=Fin	9 builds	advcl
6	something	something	PRON	Number=Sing PronType=Ind	5 ask	obj
7	,	,	PUNCT	_	5 ask	punct
8	it	it	PRON	Case=Nom Gender=Neut Number=Sing Person=3 PronType=Prs	9 builds	nsubj
9	builds	build	VERB	Mood=Ind Number=Sing Person=3 Tense=Pres VerbForm=Fin	root	root
10	an	a	DET	Definite=Ind PronType=Art	11 answer	det
11	answer	answer	NOUN	Number=Sing	9 builds	obj
12	a	a	DET	Definite=Ind PronType=Art	14 piece	det
13	little	little	ADJ	Degree=Pos	14 piece	amod
14	piece	piece	NOUN	Number=Sing	9 builds	obj
15	at	at	ADP	_	17 time	case
16	a	a	DET	Definite=Ind PronType=Art	17 time	det
17	time	time	NOUN	Number=Sing	14 piece	nmod
18	,	,	PUNCT	_	17 time	punct
19	each	each	DET	PronType=Tot	20 time	det
20	time	time	NOUN	Number=Sing	17 time	conj
21	picking	pick	VERB	VerbForm=Ger	9 builds	advcl
22	a	a	DET	Definite=Ind PronType=Art	23 word	det
23	word	word	NOUN	Number=Sing	21 picking	obj
24	that	that	PRON	PronType=Rel	25 fits	nsubj
25	fits	fit	VERB	Mood=Ind Number=Sing Person=3 Tense=Pres VerbForm=Fin	23 word	acl:relcl
26	well	well	ADV	Degree=Pos	25 fits	advmod
27	after	after	ADP	_	29 words	case
28	the	the	DET	Definite=Def PronType=Art	29 words	det
29	words	word	NOUN	Number=Plur	25 fits	obl
30	before	before	ADP	_	31 it	case
31	it	it	PRON	Case=Acc Gender=Neut Number=Sing Person=3 PronType=Prs	25 fits	obl
32	.	.	PUNCT	_	9 builds	punct

Pragmatics

form: statement
act: assert
voice: narrator
polarity: none
teacher: kept the draft

#	verb	tense, mood	who	what, whom	where	when	other	roles by
1	ask speech: ask	present	you	something				vn inquire-37.1.2 «NP V NP.recipient»: agent=Agent something=Recipient
2	build	present	chatbot	a answer		now		vn build-26.1-1 «NP.asset V NP NP»: agent=Asset answer=Beneficiary piece=Product
3	pick	gerund	chatbot not said here	a word				vn chew-39.2 «NP V NP»: agent=Agent word=Patient
4	fit	present	a word		chatbot	after the words	well property	vn funnel-9.3-1 «NP V NP PP.trajectory»: agent=Agent before it=Trajectory

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:2.

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): R5-strong ×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/2/