Datasets:

znhoughton
/

babylm-150m-ablated

text stringlengths 1 16.3k	domain stringclasses 6 values
Well it's just that, you know, a pound, or a hundred pounds today, is not the same as a hundred pounds in a year's time, or two, two years' time.	bnc_spoken
Right.	bnc_spoken
So that would be opportunity cost?	bnc_spoken
That's right, yes, that's right, so, in actual fact, this area here, right, is going to be the discounted sum of all future rural incomes.	bnc_spoken
Alright, now if we look at the, the rural instead of the urban wage rate, right, up here alright, now let's just say that it takes that amount of time before this individual gets a job in the urban area, alright, now if we discount alright the erm, the rural, the urban wages right, that's all this Why have you just ...	bnc_spoken
Sorry?	bnc_spoken
Why have you discounted it to W R?	bnc_spoken
Well it's, well, er, there's no need why that should be the case, it could be, you know, it could look something like that.	bnc_spoken
But, up in the, there the migrant's decision making process will be, is this area here right, is that area there, greater than that area there.	bnc_spoken
Alright, so this is the discounted sum in er, in non-agriculture, in the urban area, and this is the value, the discounted sum in agriculture.	bnc_spoken
Now, clearly on the, on the way I've drawn this diagram it is.	bnc_spoken
Right, but that's going to depend on right, not only the wage differential, now if the wage differential is very large it's likely that this discounted sum is going to, you know, be larger than that.	bnc_spoken
But it also depends on the time it takes to get the job.	bnc_spoken
I mean if we got a job somewhere out here then it clearly wouldn't be.	bnc_spoken
So it's this, it's the time taken to get this erm, er, to get the job's going to be important.	bnc_spoken
Now if we think of W U as being er, the expected value of the urban wage, and that equals the probability of getting a job, multiplied by the actual urban wages, alright then this probability of getting a job is going to be important as well.	bnc_spoken
And that's it then, scruffy diagram, but what more have we got?	bnc_spoken
An urban wage, a rural wage, right, we discount those wages over time, right and that area there, shading the area in the blue sort of box, then this it's erm, it's going to be rational for this er, migrant, this person to become a migrant.	bnc_spoken
Even though, I mean this doesn't say anything about unemployment levels, levels of unemployment.	bnc_spoken
Unem the level of unemployment is subsumed between, well, within the time it takes to get a job, and also the probability of, of getting a job.	bnc_spoken
Alright.	bnc_spoken
So this sort of, even that's just what the Harrison model erm, tells us, right, so that's, unemployment rates are, are virtually erm, unimportant in the migrant's decision.	bnc_spoken
Unemployment rates now, it's because they may well be, if they're acting rationally, discount over a very large, long period of time.	bnc_spoken
Okay just er, just to ask, is this erm, migration, it, does it cause a problem?	bnc_spoken
I mean I thought that migration was the er, would be the cure of all ills, in that, you know, we've got wage differentials here in this, in this hypothetical economy.	bnc_spoken
Surely if you migrate that increases the supply of labour, it reduces the supply of labour in the rural sector, it increases the supply of labour in er, the urban sector, unless wage, wage rates should equalize?	bnc_spoken
No.	bnc_spoken
What, what, why don't, why doesn't, why don't wage rates equalize?	bnc_spoken
Because erm, the urban sector is normally considered to be capital intensive a sector, and the rural area in L D Cs are considered more labour intensive, and so erm, obviously when you get an influx of labour, and changes to the capital to labour ratio Mhm erm, and you would see er, a shift in capital to, to agricultur...	bnc_spoken
However, there is a problem, capital is not perfectly mobile, like you can apply a new machine into, well maybe you could cos that's .	bnc_spoken
Whatever, you can't apply the urban type of capital back onto the land, so it's okay to say this would work if capital was completely and perfectly mobile, but it isn't so, you don't get that and you don't get balanced growth of that.	bnc_spoken
Because er, everything shifts in the wrong direction.	bnc_spoken
That's right, I mean, in order to eq equalize wages between sector, labour and also capital has to be perfectly mobile.	bnc_spoken
Now, if it's not, we know that labour isn't, isn't perfectly mobile, it appeared, migration would seem to suggest it's, it's pretty mobile geographically, but it may not be particularly mobile occupationally.	bnc_spoken
So that these people go to the urban areas to get the jobs, they're not trained erm, for these jobs right and as a result, these wages rates still, may still be maintained despite the fact there are lots of people who are quite happy to take wages erm, in, take up jobs in the urban sectors, it's the fact that sort of,...	bnc_spoken
Also, where, what are the erm, what're the highest paid jobs in the urban sector?	bnc_spoken
They are essentially in developing countries.	bnc_spoken
They are essentially government jobs.	bnc_spoken
Right.	bnc_spoken
Now, er, without wishing to generalize too much, bribery, corruption, you know, it's not what you know, it's who you know, are very important factors in er, in employment in developing countries.	bnc_spoken
And as a result they are It's much nicer to say patronage, you promise to say patronage.	bnc_spoken
Ah, patronage, okay, that's er, a nice euphemism to see the Then you don't like slander people.	bnc_spoken
That's right, in that erm, you're, the probability of you getting a well paid job is greatly enhanced, if not guaranteed if you know er, the people who are employing you.	bnc_spoken
And this sort of thing goes, is, is rife in developing countries.	bnc_spoken
I mean the old boy network is pretty bad in this country, but er, it's similar sorts of things go on in developing countries, and because there's much less of an industrial sector there, the government sector itself, erm, plays a very important role in, in employment.	bnc_spoken
Now, was it you Lyn who had some statistics some saying er, what proportion Fine I've got,I've got to use stat statistics to Well it works, erm, there were some Yes.	bnc_spoken
I can't remember if it was you or somebody else, had er, some data as to how many of the er, jobs in urban areas where the government I put erm, studies by Helen and Tate, nineteen eighty three, found public sector employment averaged forty four percent of non-agricultural employment in twenty L D Cs and in extreme cas...	bnc_spoken
And I think he quote those, or they didn't quote them in much, as much detail as I did in the lectures.	bnc_spoken
I'd got these figures down, but I hadn't really got anything against them, he just quickly went through the percentages so yes.	bnc_spoken
Okay, so And developed countries averaged twenty four percent, so it's twice as much as developed countries.	bnc_spoken
Right, so given sort of erm, public and semi-public institutions, right, represent a large proportion of the non- agricultural erm, employment opportunities, and as a result, alright, the rates of pay in the, in the civil service, essentially, are going to, going to determine erm, the urban, the urban wage rate.	bnc_spoken
Now if those, erm, wages in the civil service are set, not through sort of market process, but sort of through institutional er, constraints, you know, it's unlikely there's going to be sort of er, union power erm, in these sorts of jobs, but nevertheless, civil servants tend to do quite well at giving themselves pay i...	bnc_spoken
and er, and as a result, you know, institutes, the wages in government bodies tend to be quite high, and because gov government bodies represent such a large proportion of er, non- agricultural employment they, they represent a much more sort of, a much more potent influence than they would do in a, in a developed coun...	bnc_spoken
Erm okay what was the importance of implication of er, Harrison model?	bnc_spoken
There was one policy implication.	bnc_spoken
That, what was it, that erm people still move.	bnc_spoken
Obviously, we that right, people will still move if unemployment's high, but if you try, if the government tries and gets rid of unem urban unemployment, say people have moved because of it, so there's unemployment there to start with because of the reasons we've discussed earlier, so then the government said okay, wel...	bnc_spoken
And, erm, you won't solve anything.	bnc_spoken
That's right, so, if you want to try and erm, minimize urban unemployment, right, or it's futile to try and minimize urban unemployment by erm, er, establishing employment er, job creation schemes in urban areas, because that will increase the probability of getting a job if you're a migrant.	bnc_spoken
Right, so there are these government job creation schemes, alright.	bnc_spoken
Now, because it's typically observed in the Harrison model shows this under certain circumstances to be the case, that, erm, a sort of erm, migration elasticity with respect to income differentials, right, is much greater than the erm, The rate of unemployment, the level of unemployment That's right, yes, yes, that's ...	bnc_spoken
If job creation schemes aren't in urban areas, are the best way to get round them, isn't it, if we can show, quite simply, that job creation schemes lead to more migration, which lead to more unemployment.	bnc_spoken
The policy's self repeating, what policies might ?	bnc_spoken
A policy of cross-training,for all areas.	bnc_spoken
Okay, yeah, so, what might those, what form might those policies take, what might they be?	bnc_spoken
on it .	bnc_spoken
Yes, population controls, that's right, that's one, I mean, one reason, one way of getting rid of the migration is to, is to, is to limit population growth.	bnc_spoken
Where's the population growth highest?	bnc_spoken
It's at its highest in rural areas, this is what's being done.	bnc_spoken
Erm, population control is always focused out in the rural areas,the most difficult place to do it, because populations are dispersed, er, that's where population growth, right, is the highest.	bnc_spoken
Any other?	bnc_spoken
Making incentives to actually raise the wage rates in the, in the rural areas so, sort of to, to increase production or, and through, through farming.	bnc_spoken
To improve their situation, Right, yeah improve their wages through farming and perhaps That's right, put more infrastructure in so that then the farmers can send their children to school rather than work on the land.	bnc_spoken
yeah, that's right, what we want to try and do is to improve the erm, the returns to, to agriculture which is the main form of er, sort of income in rural areas.	bnc_spoken
I mean if you do things like that while improving infrastructure, erm, setting up credit, government credit facilities, er, so, so, so we can lend, we could lend money to small farms.	bnc_spoken
Increase in information about unem unemployment itself, that is a, that is a problem, people have this misapprehensions about the probability of getting a job You could also erm, start to recognize the benefit of the rural sector, and one reason why they were discriminating, L D Cs tended to want to ignore that and sor...	bnc_spoken
It's not going to disappear, cos it's quite a large part of urban areas.	bnc_spoken
They could do that.	bnc_spoken
Yes, we want to try and improve resource allocation in the economy and one way of doing that is not necessarily subsidizing agricultural production, but remove the taxes on agricultural growth.	bnc_spoken
Alright, and effectively what's, what's happening is that the government sectors in erm, in developing countries are very highly subsidized.	bnc_spoken
Right, they're overmanned, alright, they have very high wages, alright, er, that is one of the major courses of er, causes of resource mis-allocation, it's this very large, very inefficient government sector.	bnc_spoken
And er, one way to, to get people to stay on the land is to introduce some sort of, what we might call market disciplines, if that wasn't such a dirty word, er, into the government, the government sector where there are clear inefficiencies.	bnc_spoken
Okay, right, erm it's nearly time to go, but before we do, can I just give you some bits and pieces, you, you may well have copies of last year's exam paper, but if you haven't, this is for development and integration of trade er, have a look at those sorts of essays you're being asked to do.	bnc_spoken
But also, look as those er, those short answers, now it's important that you answer the short answer questions well alright, because it's, it's a lot easier to get good marks on a short answer question, providing you do it well, than it is on a long answer question.	bnc_spoken
Alright, erm, but most people don't, don't ans don't answer short answer questions very well at all, and that's why they get low, low marks for them.	bnc_spoken
Always re bear in mind, that whenever you do a short answer question, right, you've only got fifteen minutes to do it in.	bnc_spoken
Right, there, there isn't a great deal that you can get in, in fifteen minutes.	bnc_spoken
But what, what you must have, right, is a definition of erm, the er, sort of thing that you're asked to write about.	bnc_spoken
I mean, factor price equalization, intra-industry measuring intra-industry trade, optimal intervention, reciprocal dumping, they're all jargon, bits of jargon effect effectively, so you must, you must have a definition.	bnc_spoken
Right now, although the determinals in industrial specialization within countries, alright, now although you can't really define that, you might be able to say something about industrial specialization.	bnc_spoken
You know it's fairly obvious what it is, but with most of these will, will require, require a definition.	bnc_spoken
Right, so you must put a definition, a definition in there.	bnc_spoken
An example of it in use, and you must have ex what I'm saying, it's very important to have empirical erm, examples in your essays, right, because they will also be required in the short answer questions as well.	bnc_spoken
Right, you know, what is an example of erm, reciprocal dumping?	bnc_spoken
Right now you may, erm,don't know anything in particular, but I know that the European Community and the United, United States are always trying to out, erm, beat each other in terms of subsidies on food products, so put, put that in.	bnc_spoken
Reciprocal dumping, erm, you know, one very good example would be agricultural trade between European Community and erm, and America.	bnc_spoken
Alright, if it's a, so you, you've got to have a definition, alright, you must have a, an example of it in, in action.	bnc_spoken
Any form of empirical evidence is useful, alright, erm sometimes you know, you're asked, you know, what is the effective protection, or define, or write brief a brief essay on effective protection, right you'll, you'll want to say something there about how it's measured.	bnc_spoken
Now what essentially, what you're doing, is that you're writing an essay in very, very condensed form, so it's just, I mean you still, still must have the same recipe in there when you write a short answer, right, but you just have to be very, very brief, so you know,para paragraph on definition, you know, paragraph of...	bnc_spoken
You know, nobody's expecting three or four sides, right, you've got quarter of an hour, not erm, you know, if, if you can write a side a half you, you'll be doing much better than most people, alright, so you're just thinking about a side really, depending on how big your handwriting is.	bnc_spoken

End of preview. Expand in Data Studio

znhoughton/babylm-150m-v3 — binomial-ablated

This dataset is a copy of znhoughton/babylm-150m-v3 with sentences containing specific novel N-and-N binomial pairs removed.

Ablation details

Source corpus: znhoughton/babylm-150m-v3
Exclusion list: novel_binomials_curated.csv (431 pairs)
Sentences removed: 47
Date: 2026-04-12

Matching rule

A sentence is removed if it contains both words of an excluded pair and the word and, with and appearing between the two words in either order (any number of intervening words allowed).

Example: "cold bread and small butter" is removed for the pair (bread, butter). Sentences containing only one of the two words, or lacking and, are retained.

Domains and sentence-splitting

Domain	Structure	Ablation method
`bnc_spoken`	sentence-per-line	line-wise
`childes`	sentence-per-line	line-wise
`open_subtitles`	sentence-per-line	line-wise
`switchboard`	sentence-per-line	line-wise
`gutenberg`	multi-sentence/line	join → sentence-tokenise → filter
`simple_wiki`	multi-sentence/line	join → sentence-tokenise → filter

Excluded pairs

abashed / sorry
abbots / acolytes
abysses / puddles
acorns / boulders
acorns / goblets
acorns / lions
acorns / tigers
acorns / valleys
actresses / lumberjacks
admirals / deckhands
admirals / flowers
admirals / knitting
admirals / paupers
admirals / ribbons
admirals / weavers
allergic / unaccustomed
amber / gravel
amethysts / pebbles
amphitheaters / gazebos
annoying / teal
anthills / volcanoes
ants / giraffes
ants / stars
anvils / coins
aprons / swords
archbishops / millers
archdukes / tanners
archdukes / vagrants
archipelagos / sandbars
arenas / sheds
arrows / flowers
asteroids / pebbles
avalanches / feathers
avalanches / murmurs
avalanches / snowballs
axes / flowers
axes / petals
axes / ribbons
bacteria / candy
ballrooms / cellars
banners / rags
banquets / crumbs
banyans / ferns
barns / marbles
barons / fishermen
barons / paupers
barons / weavers
bassoons / kazoos
battleaxes / penknives
battlements / cobwebs
bears / buttons
bears / kettles
bears / ribbons
bears / seeds
bears / teacups
beautiful / stinky
beetles / hills
beetles / mammoths
beetles / mastodons
beetles / moons
beggars / captains
beggars / dukes
bicycles / robots
birdhouses / cathedrals
birdhouses / skyscrapers
birdhouses / turrets
bishops / blacksmiths
bishops / potters
bishops / seamstresses
bishops / vagabonds
bison / sparrows
bitter / purple
blacksmiths / lawyers
blankets / kittens
blizzards / butterflies
blizzards / flurries
blizzards / mists
bogs / tundras
bombs / pebbles
bonnets / cannons
bonnets / shields
bonnets / swords
bonnets / wolves
books / owls
boulders / butterflies
boulders / feathers
boulders / marbles
bowls / sharks
breezes / tornadoes
breezes / typhoons
bridges / needles
buffaloes / sparrows
burlap / cashmere
burlap / velvet
butterflies / comets
butterflies / earthquakes
butterflies / floods
butterflies / squalls
butterflies / tempests
buttons / castles
buttons / deer
buttons / foxes
buttons / lions
buttons / mountains
buttons / seas
buttons / whales
cabins / citadels
campfires / wildfires
candles / foxes
candles / galaxies
candles / gales
candles / hurricanes
candles / nebulae
cannons / darts
cannons / needles
cannons / slingshots
canyons / marbles
captains / flowers
cardinals / novices
carracks / skiffs
cashmere / flannel
catapults / slingshots
cathedrals / outhouses
cathedrals / sheds
cauldrons / thimbles
cedars / mosses
cellos / kazoos
cellos / triangles
chalices / thimbles
chalk / diamonds
chalk / tanzanite
champagne / vinegar
chandeliers / torches
chanting / enchanting
chasms / puddles
chauffeurs / stewardesses
cherries / llamas
chestnuts / ferns
chickens / fences
chipmunks / gorillas
churches / pebbles
civilizations / homesteads
clarinets / kazoos
clay / emeralds
claymores / penknives
clearings / plateaus
clearings / prairies
cliffs / cobwebs
cliffs / moths
clocks / deer
closets / dungeons
cobblers / colonels
cobblers / deacons
cobblers / judges
cobblers / marquesses
cobblers / monks
cobblers / privateers
cobras / mice
cobwebs / storms
coins / eagles
coins / forests
coins / ravens
comets / sparks
condors / sparrows
conscripts / generals
coroners / senators
corsairs / drapers
cottages / dynasties
counts / shepherds
counts / tanners
counts / vagabonds
crabs / walruses
crates / warehouses
crayfish / otters
crocodiles / tadpoles
crossbows / slingshots
crowbars / tweezers
crowns / nightcaps
crumbs / feasts
cudgels / maces
cutlasses / needles
cyclones / gusts
cyclones / zephyrs
cypresses / mosses
daggers / flowers
daggers / thimbles
deltas / puddles
deserts / raindrops
determined / forgettable
dewdrops / forests
dewdrops / glaciers
dewdrops / planets
dewdrops / tsunamis
dewdrops / waterfalls
disheveled / dreary
ditches / fjords
doctors / shepherds
donates / provides
dreadnoughts / rowboats
drips / geysers
dukes / flowers
dukes / millers
dukes / shepherds
dukes / vagrants
dungeons / spiders
eagles / plates
eagles / ribbons
eagles / thimbles
earls / gravediggers
earls / paupers
earls / tanners
earthquakes / teardrops
earthquakes / whimpers
elephants / gnats
elephants / pins
elms / mushrooms
embers / thunderbolts
emperors / flowers
eons / heartbeats
eruptions / sparks
falcons / wrens
farmers / princes
farmers / shoguns
feasts / morsels
feasts / scraps
feathers / floods
feathers / lightning
feathers / mountains
feathers / rivers
feathers / spears
feathers / storms
feathers / thunderstorms
feathers / tsunamis
feathers / volcanoes
felines / quails
ferns / larches
ferns / sequoias
ferns / spruces
ferns / walnuts
finches / vultures
fireflies / galaxies
fireflies / moons
fireflies / supernovas
first / ninety-eighth
fishermen / generals
flails / switches
flint / opals
flintlocks / slingshots
floods / puddles
flowers / generals
flowers / knights
flowers / knives
flowers / lances
flowers / lords
flowers / monarchs
flowers / pharaohs
flowers / princes
flowers / warriors
flowers / zinnias
flutes / kazoos
footmen / marshals
fortresses / tents
foxes / thimbles
friars / tanners
fuming / mad
gales / sparks
galleons / kayaks
garnets / slate
gazelles / leopards
gelatin / lard
generals / knitters
generals / lace
generals / riflemen
gerbils / pythons
glaciers / raindrops
gnats / gorillas
gnats / hippopotamuses
gnats / mammoths
gnats / rhinoceroses
goblets / puddles
gorillas / sparrows
gravediggers / kings
gravel / moonstone
gravel / onyx
gravel / sapphires
gravel / turquoise
groundskeeper / superintendent
hailstones / icebergs
hairpins / sabres
hairpins / tiaras
halberds / knives
hammers / lace
happily / rudely
harpooners / knitters
harps / kazoos
harpsichords / kazoos
hawks / thimbles
hawks / voles
helmets / nightcaps
helmets / ribbons
herrings / seals
herrings / sharks
hesitate / readjust
hippopotamuses / mice
horses / mirrors
horses / thimbles
hurricanes / petals
hurricanes / zephyrs
huts / monasteries
icebergs / snowflakes
jacket / phone
jaguars / rabbits
kaisers / peasants
kale / vegetables
kazoos / lutes
kazoos / organs
kazoos / pianos
kazoos / saxophones
kazoos / trumpets
kazoos / violas
kazoos / violins
khans / weavers
lace / twine
lances / scarves
lanterns / ravens
laundresses / queens
leopards / rabbits
leviathans / minnows
lightning / moths
lions / needles
lions / teacups
lizards / tyrannosaurs
looms / trebuchets
lords / shepherds
maelstroms / ripples
majors / spinners
mansions / pigsties
mantles / rags
marbles / orbs
marooned / missing
masculine / undignified
mastodons / moths
mausoleums / shacks
meadows / tundras
meteors / raindrops
mice / tigers
mice / wolves
millers / priests
millstones / pebbles
minnows / orcas
moles / polecats
monarchs / serfs
monks / tanners
monks / telescopes
moths / seas
moths / suns
mountains / petals
mud / platinum
mudslides / puddles
mushrooms / oaks
mushrooms / pines
muskets / slingshots
nebulae / sparks
nectar / swill
needles / ravens
needles / sledgehammers
needles / tigers
nurses / patriarchs
oboes / tambourines
organs / whistles
owls / pennies
owls / thimbles
paintbrushes / sailors
palaces / pigsties
pearls / silt
peasants / sultans
pebbles / volcanoes
pelicans / sardines
pennies / wolves
petals / tornadoes
pharaohs / potters
pianos / triangles
pickaxes / pins
pikes / slingshots
pinnacles / vales
pins / tigers
pins / walls
planets / sparks
priests / tanners
priors / scullions
pterodactyls / sparrows
puddles / straits
puppies / tigers
pyramids / sandcastles
queens / serfs
rags / tapestries
rapiers / thimbles
rats / sharks
redwoods / toadstools
rhinoceroses / voles
ribbons / sharks
ribbons / shields
ripples / waterspouts
rivers / seeds
rubies / sandstone
rustles / tempests
sackcloth / velvet
sardines / whales
saxophones / whistles
scepters / twigs
scissors / scythes
seeds / sharks
serfs / tsars
sharks / teacups
showers / typhoons
slings / trebuchets
soldiers / spools
sparks / thunderbolts
sparrows / wolves
spears / tablecloths
spoons / wolves
tablecloths / tigers
teacups / wolves
therapy / vacations
thimbles / towers
thimbles / wolves
thunderclaps / whispers
trombones / whistles
tsars / vagrants
tubas / whistles
vocabulary / vowels

Downloads last month: 19

Models trained or fine-tuned on znhoughton/babylm-150m-ablated