Mova0.5 pre-release⌕

☰

home / corpora / questions / AI / How a language model learns: guessing… / t6

Test 6 · How a language model learns: guessing the next word

Ask ten bots for a number from 1 to 10. Which is most likely?

Mova's choice: option 1 · The right answer: option 1 · right ✓ · Who acts: you · What happens: ask, be · To what, where: ten bots

This is test 6 of lesson 2 «How a language model learns: guessing the next word» of the AI course of Mova's primary school. The page gives the question's syntax tree — a dependency tree in the style of Universal Dependencies — and its pragmatics clause by clause, then the 4 options as the student saw them, each with its tree and pragmatics. Every text is parsed by Mova itself. How the student answered it, with the score of every option and why: the test at school.

The question

Sentence 1: Ask ten bots for a number from 1 to 10.

The words, one row each (CoNLL-U)

#	word	lemma	UPOS	features	head	relation
1	Ask	ask	VERB	Mood=Imp VerbForm=Fin	root	root
2	ten	ten	NUM	NumForm=Word NumType=Card	3 bots	nummod
3	bots	bot	NOUN	Number=Plur	1 Ask	obj
4	for	for	ADP	_	6 number	case
5	a	a	DET	Definite=Ind PronType=Art	6 number	det
6	number	number	NOUN	Number=Sing	1 Ask	obl
7	from	from	ADP	_	8 1	case
8	1	1	NUM	NumForm=Digit NumType=Card	6 number	nmod
9	to	to	ADP	_	10 10	case
10	10	10	NUM	NumForm=Digit NumType=Card	8 1	nmod
11	.	.	PUNCT	_	1 Ask	punct

#	verb	tense, mood	who	what, whom	to whom	where	when	other	roles by
1	ask speech: ask	imperative	you not said here	ten bots				for a number from 1 to 10	vn inquire-37.1.2 «NP V NP.recipient»: agent=Agent bot=Recipient

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.

Sentence 2: Which is most likely?

The words, one row each (CoNLL-U)

#	word	lemma	UPOS	features	head	relation
1	Which	which	PRON	PronType=Rel	4 likely	nsubj
2	is	be	AUX	Mood=Ind Number=Sing Person=3 Tense=Pres VerbForm=Fin	4 likely	cop
3	most	most	ADV	Degree=Sup	4 likely	advmod
4	likely	likely	ADV	_	root	root
5	?	?	PUNCT	_	4 likely	punct

#	verb	tense, mood	who	what, whom	to whom	where	when	other	roles by
1	is	present						likely property	tree rules

The options

Option 1 · ← Mova's choice · ✓ the right answer

Many of them give 7, like many people do

The words, one row each (CoNLL-U)

#	word	lemma	UPOS	features	head	relation
1	Many	many	ADJ	Degree=Pos	4 give	nsubj
2	of	of	ADP	_	3 them	case
3	them	they	PRON	Case=Acc Number=Plur Person=3 PronType=Prs	1 Many	nmod
4	give	give	VERB	Mood=Ind Number=Plur Person=3 Tense=Pres VerbForm=Fin	root	root
5	7	7	NUM	NumForm=Digit NumType=Card	4 give	obj
6	,	,	PUNCT	_	10 do	punct
7	like	like	SCONJ	_	10 do	mark
8	many	many	ADJ	Degree=Pos	9 people	amod
9	people	people	NOUN	Number=Plur	10 do	nsubj
10	do	do	VERB	Mood=Ind Number=Plur Person=3 Tense=Pres VerbForm=Fin	4 give	advcl

#	verb	tense, mood	who	what, whom	to whom	where	when	other	roles by
1	give	present	Many of them	7					vn give-13.1-1 «NP V NP»: agent=Agent 7=Theme
2	do	present	many people						vn act-114-1 «NP V»: agent=Agent

Option 2

Each gives a different number, fairly spread out

The words, one row each (CoNLL-U)

#	word	lemma	UPOS	features	head	relation
1	Each	each	DET	PronType=Tot	2 gives	nsubj
2	gives	give	VERB	Mood=Ind Number=Sing Person=3 Tense=Pres VerbForm=Fin	root	root
3	a	a	DET	Definite=Ind PronType=Art	5 number	det
4	different	different	ADJ	Degree=Pos	5 number	amod
5	number	number	NOUN	Number=Sing	2 gives	obj
6	,	,	PUNCT	_	8 spread	punct
7	fairly	fairly	ADV	_	8 spread	advmod
8	spread	spread	VERB	Tense=Past VerbForm=Part Voice=Pass	2 gives	conj
9	out	out	ADP	_	8 spread	compound:prt

#	verb	tense, mood	who	what, whom	to whom	where	when	other	roles by
1	give	present	Each	a different number					vn give-13.1-1 «NP V NP»: agent=Agent number=Theme
2	spread out	participle	Each not said here					fairly property	vn appear-48.1.1 «NP V»: agent=Theme

Option 3

They all say 10, because it is the biggest one

The words, one row each (CoNLL-U)

#	word	lemma	UPOS	features	head	relation
1	They	they	PRON	Case=Nom Number=Plur Person=3 PronType=Prs	3 say	nsubj
2	all	all	DET	PronType=Tot	3 say	advmod
3	say	say	VERB	Mood=Ind Number=Plur Person=3 Tense=Pres VerbForm=Fin	root	root
4	10	10	NUM	NumForm=Digit NumType=Card	3 say	obj
5	,	,	PUNCT	_	11 one	punct
6	because	because	SCONJ	_	11 one	mark
7	it	it	PRON	Case=Nom Gender=Neut Number=Sing Person=3 PronType=Prs	11 one	nsubj
8	is	be	AUX	Mood=Ind Number=Sing Person=3 Tense=Pres VerbForm=Fin	11 one	cop
9	the	the	DET	Definite=Def PronType=Art	11 one	det
10	biggest	big	ADJ	Degree=Sup	11 one	amod
11	one	one	NOUN	Number=Sing	3 say	conj

#	verb	tense, mood	who	what, whom	to whom	where	when	other	roles by
1	say speech: assert	present	They	10					vn say-37.7-1-1 «NP V NP.topic»: agent=Agent 10=Topic
2	is a	present	it	the biggest one					tree rules

Option 4

They all politely refuse to pick any number

The words, one row each (CoNLL-U)

#	word	lemma	UPOS	features	head	relation
1	They	they	PRON	Case=Nom Number=Plur Person=3 PronType=Prs	4 refuse	nsubj
2	all	all	DET	PronType=Tot	1 They	det
3	politely	politely	ADV	_	4 refuse	advmod
4	refuse	refuse	VERB	Mood=Ind Number=Plur Person=3 Tense=Pres VerbForm=Fin	root	root
5	to	to	PART	_	6 pick	mark
6	pick	pick	VERB	VerbForm=Inf	4 refuse	xcomp
7	any	any	DET	PronType=Ind	8 number	det
8	number	number	NOUN	Number=Sing	6 pick	obj

#	verb	tense, mood	who	what, whom	to whom	where	when	other	roles by
1	refuse	present	They					politely property	vn reject-77.2 «NP V S_ING»: agent=Agent clause=Theme
2	not pick	to-infinitive	They not said here	any number					vn chew-39.2 «NP V NP»: agent=Agent number=Patient

Text: the tests of Mova's primary school, written for Mova (world/data/school/), as the course runner gave them to the student. Trees: Mova's own parse — the en model on the CPU, when the site is built — not checked by a teacher; they may be wrong. Frames: Mova's frames reader (prag), each text read on its own. Elamur Cosmic License.

This page in Markdown → https://md.mova.elamur.ai/free/en/corpora/questions/ai/2-guessing-the-next-word/t6/