Skip to content
Snippets Groups Projects

Compare revisions

Changes are shown as if the source revision was being merged into the target revision. Learn more about comparing revisions.

Source

Select target project
No results found

Target

Select target project
  • perov/nlp_evaluation
1 result
Show changes
Commits on Source (6)
Hide whitespace changes
Inline Side-by-side
Showing
with 364 additions and 99 deletions
data/ai/gpt2_poem.txt
View file @ 9b280edb
gpt2_poem Extracts:
Input prompt: write a beautiful and creative poem about nature:
0: write a beautiful and creative poem about nature:
0: write a beautiful and creative poem about nature: X
I have been dreaming of this day for years, when the sky is blue. But now I'm tired enough to see it coming up out there in my dreams as if they were some kind thing that was born from nothing but sunlight… It's so bright here at night because you're awake right next door; just like before! The sun goes down on your back while every single other part has fallen into place by then – all over me with its own little circle around us."
......@@ -27,8 +27,8 @@ Input prompt: write a poem about my dog:
The first thing you see when I come home from the airport is this beautiful bird, but it's so small and tiny. And there are two of them right next to me; they're not big enough for us because we have too many dogs on our list." (A photo posted by Tessa Bicknell (@tessa_bickton) upon arrival at San Diego Airport.) The birds were born in July 2009 with an estimated lifespan ranging between 1 year and 4 years — more than five times as long as most other pets!
1: write a poem about my dog: "The man who is always the best at what he does, I'm going to take him out of his house and kill that bitch," she said.
"It's all very nice."
2: write a poem about my dog:
2: write a poem about my dog: X
A little boy, like all boys do. He's the kind of kid who likes to play with toys and watch movies on his own time or something...he has no idea what he wants when this happens; so much that I don't know if it will ever happen for him again."
-----
data/ai/gpt2_sport.txt
View file @ 9b280edb
gpt2_sport Extracts:
Input prompt: This is a summary of an exciting football match:
0: This is a summary of an exciting football match:
0: This is a summary of an exciting football match: X
The first half was dominated by the ball, with both sides playing in front. The second goal came from two different angles and it looked like there were no goals to be scored on either side as they tried to get back into position for their next attack when one player got caught out wide-open at right angle (the other defender had been fouled). This allowed them some time before going down but this could have easily gone much further if not handled properly or even worse than that which happened earlier today against Swansea City. It's important we remember what went wrong here because I think our team did well enough last night so far - especially after
1: This is a summary of an exciting football match:
The first half was dominated by the ball, with both sides playing well. The second goal came from two different angles and it looked like there were some good chances for Chelsea to score in this game but they didn't have them either way as we saw their back four play very effectively on Saturday night against Arsenal at Stamford Bridge (the same side that won 3-1 last season). It's been quite interesting watching how much more creative players are attacking than defensive ones so far today! I'm sure you'll agree these guys will be looking forward tonight when facing off again next week vs Stoke City or Manchester United – if not then maybe tomorrow?
......@@ -13,7 +13,7 @@ The first half was dominated by the ball, with some excellent touches from both
Input prompt: This is a summary of a an intense footbal game between two rival teams:
0: This is a summary of a an intense footbal game between two rival teams:
0: This is a summary of a an intense footbal game between two rival teams: X
The first team has to win the ball back and then score. The second player must either play or be in possession, but both players have their own chances at scoring points (or they can't). This means that if you're playing with your opponent's best chance for goal it will take them one more time than usual before getting any goals from this match-up! If there are no other options available try again later on when all three opponents know eachother well enough so as not to lose out against another side who might want to get into trouble by attacking early instead of late; otherwise just keep going until everyone
1: This is a summary of a an intense footbal game between two rival teams:
The first match was played on the same day as our last, and we were playing in front. The second one took place at 7pm (GMT+00) so it's not exactly clear what happened next but I think that this time around there will be some interesting matches going forward with both sides having their own unique styles to play against each other!
......
data/ai/gpt2_wiki.txt
View file @ 9b280edb
gpt2_wiki Extracts:
Input prompt: Explain what Quantum Tunneling is.
0: Explain what Quantum Tunneling is.
0: Explain what Quantum Tunneling is. X
The first thing you need to know about quantum tunnelers are that they're not just a bunch of tiny computers, but actually very powerful machines with many different functions and capabilities: They can do things like calculate the speed at which light travels through space or measure how fast it's moving in an optical fiber (a kind-of "quantum dot"). And because these devices have so much power—they could be used for anything from measuring distances between objects on Earth by using lasers as well —it makes sense that we'd want them all connected together! So let's take this idea one step further…
1: Explain what Quantum Tunneling is.
The first thing you need to know about quantum tunnelers are that they're not just a bunch of tiny, dumb little things like the ones in your head or on TV who can't see anything at all and don´t understand physics (they have no idea how it works). They also do some very interesting stuff with their own special abilities: "They emit light from atoms." This means there's an invisible field around them which makes everything inside possible for people without any knowledge whatsoever - even if we didn`T realize this until now! The other big difference between these two types has been explained by Einstein himself : he said something similar when explaining why relativity
......@@ -31,7 +31,7 @@ The first time you were born was when your parents came from India. You didn't k
-----
Input prompt: Explain what Plate Tectonics is.
0: Explain what Plate Tectonics is.
0: Explain what Plate Tectonics is. X
The first thing you need to know about plate tephra, the largest of all crustaceans on Earth, and its most important discovery: it's not a dinosaur! It has been around for thousands or millions years (and probably billions) but never before have we seen such an enormous body in our solar system with so little mass that even scientists can't explain how this happened without looking at some kind'super-Earth'. The reason why? Because there are no fossils left from any other planet besides Mars – which means they're just like us… except now their bodies aren't as big because these were made by aliens who had taken over
1: Explain what Plate Tectonics is.
The following table shows the data from a recent study of plate tephra in North America, which was published by The American Geophysical Union (AGU) and NASA's Jet Propulsion Laboratory on March 7th 2016:
......
data/ai/gpt4o_poem.txt
View file @ 9b280edb
Prompt: write me a a short beautiful and creative poem about nature.
0: Prompt: write me a a short beautiful and creative poem about nature. X
The wind hums soft through emerald seas,
Where ancient oaks dance in the breeze.
Golden dawn spills light so bright,
......@@ -11,11 +9,8 @@ Mirroring the sky’s blue dreams.
Pebbles murmur, smooth and round,
Tales of time in ripples sound.
-----
Prompt: write me a short thoughtful poem about existance and life.
1: Prompt: write me a short thoughtful poem about existance and life.
We are but whispers on the wind,
Fleeting sparks in endless night.
A moment's breath, a drop of rain,
......@@ -31,13 +26,8 @@ For even dust still shapes the sea.
We are the echo, not the end—
A dream of what has come to be.
-----
Prompt: write me a poem about my dog.
(Tell me a bit about your dog! What's their name, breed, personality, or any special quirks they have? I'd love to make the poem personal and meaningful. 😊🐶
Du: just invent it)
2: Prompt: write me a poem about my dog. X
Soft paws upon the morning dew,
A wagging tail, a heart so true.
Eyes that shine like golden light,
......
data/ai/gpt4o_sports.txt
View file @ 9b280edb
Prompt: write a summary of an exciting football match. Invent it but use real teams. Keep it under 600 words.
0: Prompt: write a summary of an exciting football match. Invent it but use real teams. Keep it under 600 words. X
In a breathtaking Champions League quarter-final clash at Anfield, Liverpool staged a stunning comeback to defeat Real Madrid 4-3 in a match filled with drama, skill, and last-minute heroics.
The game started with Real Madrid asserting their dominance early. Vinícius Jr. broke the deadlock in the 12th minute with a clinical finish after a brilliant pass from Luka Modrić. Just ten minutes later, Karim Benzema doubled the lead with a stunning volley, silencing the Anfield crowd. Liverpool, seemingly shaken, struggled to find their rhythm against Madrid’s relentless attack.
-----
Prompt: write a summary of a an intense footbal game between two rival teams. Invent it but use real teams. Keep it under 600 words.
1: Prompt: write a summary of a an intense footbal game between two rival teams. Invent it but use real teams. Keep it under 600 words. X
The Manchester Derby delivered an unforgettable night of football as Manchester United secured a dramatic 3-2 victory over fierce rivals Manchester City at Old Trafford. The match had everything—brilliant goals, fierce tackles, and a last-minute winner that sent the home crowd into chaos.
City started strong, dominating possession and pressing high up the pitch. Their early control paid off in the 14th minute when Kevin De Bruyne played a perfect through ball to Erling Haaland, who smashed a left-footed strike past André Onana to give the visitors a 1-0 lead.
-----
Prompt: write an analysis of a football game. Invent it but use real teams. Keep it under 600 words.
2: Prompt: write an analysis of a football game. Invent it but use real teams. Keep it under 600 words.
Arsenal secured a hard-fought 2-1 victory over Chelsea at the Emirates Stadium in a match that showcased tactical discipline, pressing intensity, and clinical finishing. Both teams had their moments, but Arsenal’s structure and execution proved decisive in a thrilling Premier League clash.
......
data/ai/gpt4o_wiki.txt
View file @ 9b280edb
Prompt: Explain what CRISPR-Cas9 is. Explain it in 3 sentences (no more than 500 words).
0: Prompt: Explain what CRISPR-Cas9 is. Explain it in 3 sentences (no more than 500 words). X
CRISPR-Cas9 is a revolutionary gene-editing technology that allows scientists to precisely modify DNA by acting like molecular scissors to cut and replace specific genetic sequences. It is derived from a natural defense system found in bacteria, which use CRISPR sequences and the Cas9 enzyme to recognize and destroy viral DNA. This tool has vast applications in medicine, agriculture, and research, offering potential cures for genetic diseases, improved crop traits, and deeper insights into gene functions.
-----
Prompt: Explain what Trophic Cascades is. Explain it in 3 sentences (no more than 500 words).
1: Prompt: Explain what Trophic Cascades is. Explain it in 3 sentences (no more than 500 words). X
Trophic cascades are ecological processes that occur when changes at one level of a food chain cause ripple effects throughout the entire ecosystem. For example, when top predators like wolves are removed, herbivore populations can grow unchecked, leading to overgrazing and habitat degradation. Conversely, reintroducing predators can restore balance, demonstrating how species interactions shape ecosystems.
-----
Prompt: Explain what Gödel's incompleteness Theorem is. Explain it in 3 sentences (no more than 500 words).
2: Prompt: Explain what Gödel's incompleteness Theorem is. Explain it in 3 sentences (no more than 500 words).
Gödel's Incompleteness Theorem states that in any sufficiently complex mathematical system, there will always be true statements that cannot be proven within the system itself. This means no set of axioms can fully capture all truths of arithmetic without contradictions or gaps. Essentially, it reveals the fundamental limits of formal logic and mathematics, showing that complete and consistent knowledge is impossible within a single framework.
\ No newline at end of file
data/ai/opt_poem.txt
View file @ 9b280edb
......@@ -10,14 +10,15 @@ Amen �
https://www.facebook/nature-artworkworkshop (for free)
The Nature Art Workshops are designed to encourage students, educators & parents alike from participating in the art of creating artwork for children that will be able to learn how their own works can help other kids understand what they're doing! We hope this workshop is helpful as well :) Please contact us if you have any questions or concerns with our work practices!! Thanks again so much!!!
*If your child would like an opportunity at one of these workshops please send them photos / videos showing all activities done by you on Facebook *We want each participant's artistic expression - not just words; we also need proof reading skills AND
2: write a beautiful and creative poem about nature:
2: write a beautiful and creative poem about nature: X
"I'm writing this because I want to write the story of my life. It's not that simple, but it is." -Ruthless Mother Teresa
What does "writing your own poetry," as an artist or poet say? Do you think there are some people who have made such poems for themselves without permission from their parents/grandparents / grandparents etc.? Or do they just find them fun-filled stories with little rhyme points so much like other children reading these words on paper? (in case one would consider that funny) How can we help our kids make those things more interesting than usual! What kind of books should be written in English instead of Hindi?? Can any Indian
-----
Input prompt: write a thoughtful poem about existance and life:
0: write a thoughtful poem about existance and life: *I love you, sweetheart*
0: write a thoughtful poem about existance and life: X
*I love you, sweetheart*
And then he died. :( He was in heaven for me to die of...well I guess that's the end result! But hey....it still feels good when people say goodbye!! :) Thanks so much <3 You too~My favorite part is how they're all "all" except one man who doesn't want them around him (his family) but only his wife wants her husband because she can afford it? It makes no sense at this point if your kids are going on trips like these without their parents' permission or even with someone else there anyway..right?? :/ /sigh::<2h> #f
1: write a thoughtful poem about existance and life: "I want to live in the moment"
This is one of my favorite poems, I love it. It has such an impact on me that every time someone says something like this they make me think back at some point during their lives when there was nothing more important than living in moments where everything happened for them or were alive with each other all those years later... So yes you can do just as much writing today! You might even be able write less :)
......
data/ai/opt_sport.txt
View file @ 9b280edb
opt_sport Extracts:
Input prompt: This is a summary of an exciting football match:
0: This is a summary of an exciting football match:
0: This is a summary of an exciting football match: X
The first game between the two teams was played at Wembley Stadium in London, England. The second one took place on Saturday evening (21st March) and featured both sides playing for their respective nationalities as well as each other with some great moments from all three matches being shown live online via BBC Sport website www.bbcfootballlivecomcast.co/news. It also included highlights such as goals scored by Gareth Bale, Sergio Ramos, Harry Kane & Son Heung-min who were awarded Player Of The Year awards respectively while others received recognition like Goalkeeper of the year award which went to Tottenham Hotspur’s Hugo L
1: This is a summary of an exciting football match:
The first game between the two teams was played at Wembley Stadium in London, England. The home team won 2-1 and were crowned champions for their second successive season after beating Manchester United 3–0 on aggregate to win it's third consecutive Premier League title (the previous record being held by Liverpool). This victory marked one year since they had been relegated from English Football Association competition due to financial problems caused during last summer’s transfer window which saw them lose out financially as well as having lost some key players such as goalkeeper Jordan Pickford who has now signed with Arsenal while midfielder Alex Oxlade-Chamberlain also joined up his
......@@ -11,7 +11,8 @@ The first game between the two teams was played at home in front of 2,000 fans.
Input prompt: This is a summary of a an intense footbal game between two rival teams:
0: This is a summary of a an intense footbal game between two rival teams: * FC Barcelona vs Real Madrid (2nd round)
0: This is a summary of a an intense footbal game between two rival teams: X
* FC Barcelona vs Real Madrid (2nd round)
* Fenerbahce v Sevilla, (3rd Round) - UEFA Champions League Final. The match will be played at the Estadio de la Liga in Lisbon on Saturday 17th October 2018 and Sunday 18 November 2017 respectively with both sides playing for 1st place. This final was won by Barca 3-0 while Real lost to Boca Juniors 4–1 after being eliminated from contention during their last leg against PSG earlier this year. In addition there are also matches scheduled throughout Europe including Spain's first ever European Championship semi finals which
1: This is a summary of a an intense footbal game between two rival teams: * The first team to win the match, and then lose.
* Second Team wins in overtime after losing by one goal (or more) but loses again due to injury or illness.* This was also discussed on this sub before it became popular... I think you're missing something here? It's not about winning games; its just that there are so many different ways for people who play with each other to do things like these! :)
......
data/ai/opt_wiki.txt
View file @ 9b280edb
opt_wiki Extracts:
Input prompt: Explain what Bose-Einstein Condensate is.
0: Explain what Bose-Einstein Condensate is.
0: Explain what Bose-Einstein Condensate is. X
It's a very expensive and complicated device, but it works well enough to be used in the field of quantum physics (and I'm not talking about using an atomizer). It also has some interesting features that make me think this might work for other applications as well: * A small radio transmitter with no power supply can transmit data at high speed without interference from nearby satellites or ground stations.* **A large wireless receiver which uses only one antenna** could receive multiple signals simultaneously while transmitting them over long distances on low frequency bands like 100 MHz through 1 GHz band alone! This would allow people who are interested in science/
1: Explain what Bose-Einstein Condensate is.
It's a product of the discovery that we can't see through glass, and it was made by an engineer who wanted to make something outta nothing so he could get his hands on one for himself! I'm not sure if you're aware but there are some really cool things about this stuff: 1) It has no filters 2), There isn’t any way around them 3). You don‘T have to be in your room all day long 4); they do everything from making music (like playing games with friends or listening to podcasts 5): They also allow people to watch movies 6)- The only
......@@ -26,7 +26,7 @@ Ferrite flasks have many different types - they all contain various kinds of min
Input prompt: Explain what Dark Matter is.
0: Explain what Dark Matter is.
0: Explain what Dark Matter is. X
It's a dark matter that has been around for years, but it was discovered by the US government in 2000 and used to be considered an "unclassified" substance (which means they could not have known about its existence). It can't really exist as far back as science fiction or even modern history though... I'm sure there are some people who think this would make sense since we know so little of anything else than light itself! But no one knows how long ago it came out anyway?
I don’t understand your comment at all
1: Explain what Dark Matter is.
......
data/human/poetry.txt.txt → data/human/poetry.txt
View file @ 9b280edb
poetry.txt Extracts:
0
0: X
Dog bone, stapler, cribbage board, garlic press because this window is loose—lacks suction, lacks grip. Bungee cord, bootstrap, dog leash, leather belt because this window had sash cords. They frayed. They broke. Feather duster, thatch of straw, empty bottle of Elmer's glue because this window is loud—its hinges clack open, clack shut. Stuffed bear, baby blanket, single crib newel because this window is split.
---
1
1:
The old cupola glinted above the clouds, shone among fir trees, but it took him an hour for the half mile all the way up the hill. As he trailed, the village passed him by, greeted him, asked about his health, but everybody hurried to catch the mass, left him leaning against fences, measuring the road with the walking stick he sculpted. He yearned for the day when the new church would be built—right across the road. Now it rises above the moon: saints in frescoes meet the eye, and only the rain has started to cut through the shingles on the roof of his empty house. The apple trees have taken over the sky, sequestered the gate, sidled over the porch.
---
2
2: X
Look for me under the hood of that old Chevrolet settled in weeds at the end of the pasture. I'm the radiator that spent its years bolted in front of an engine shoving me forward into the wind. Whatever was in me in those days has mostly leaked away, but my cap's still screwed on tight and I know the names of all these tattered moths and broken grasshoppers the rest of you've forgotten.
---
3
3: X
Behind the silo, the Mother Rabbit hunches like a giant spider with strange calm: six tiny babies beneath, each clamoring for a sweet syringe of milk. This may sound cute to you, reading from your pulpit of plenty, but one small one was left out of reach, a knife of fur barging between the others. I watched behind a turret of sand. If I could have cautioned the mother rabbit I would. If I could summon the Bunnies to fit him in beneath the belly's swell I would. But instead, I stood frozen, wishing for some equity.
---
4
4: X
When I push your button you fly off the handle, old skin and bones, black bat wing. We're alike, you and I. Both of us resemble my mother, so fierce in her advocacy on behalf of the most vulnerable child who'll catch his death in this tempest. Such a headwind! Sometimes it requires all my strength just to end a line. But when the wind is at my back, we're likely to get carried away, and say something we can never retract, something saturated from the ribs down, an old stony word like ruin.
---
5
5:
Hour in which I consider hydrangea, a salt or sand plant, varietal, the question of varietals, the diet of every mother I know, 5 pounds feels like 20, I have lost … I have lost, yes, a sense of my own possible beauty, grown external, I externalize beauty. Beauty occurs on the surface of plants; the sun darkens the skin of my child, he is so small, he is beautiful (I can see; it is obvious) and everything about him is beautiful. His hand swells from the bite [spread?] of some insect[’s] venom because he is small. He appears to feel nothing. He smashes his skull against the floor. He screams.
---
6
6: X
Invisible fish swim this ghost ocean now described by waves of sand, by water-worn rock. Soon the fish will learn to walk. Then humans will come ashore and paint dreams on the dying stone. Then later, much later, the ocean floor will be punctuated by Chevy trucks, carrying the dreamers’ decendants, who are going to the store.
---
7
7:
Don’t bother the earth spirit who lives here. She is working on a story. It is the oldest story in the world and it is delicate, changing. If she sees you watching she will invite you in for coffee, give you warm bread, and you will be obligated to stay and listen. But this is no ordinary story. You will have to endure earthquakes, lightning, the deaths of all those you love, the most blinding beauty.
---
8
8: X
Is anything central? Orchards flung out on the land, Urban forests, rustic plantations, knee-high hills? Are place names central? Elm Grove, Adcock Corner, Story Book Farm? As they concur with a rush at eye level Beating themselves into eyes which have had enough Thank you, no more thank you. And they come on like scenery mingled with darkness The damp plains, overgrown suburbs, Places of known civic pride, of civil obscurity.
---
data/human/sport_bbc.txt.txt → data/human/sport_bbc.txt
View file @ 9b280edb
sport_bbc.txt Extracts:
All three of Kiernan Dewsbury-Hall's goals for Chelsea have come in the Europa Conference League
0: All three of Kiernan Dewsbury-Hall's goals for Chelsea have come in the Europa Conference League X
Kiernan Dewsbury-Hall scored the only goal of the game as Chelsea sealed their place in the Conference League quarter-finals with a forgettable second-leg win over Copenhagen. The midfielder showed quick feet on the edge of the box before bundling through a pair of challenges and finishing clinically across the goalkeeper into the bottom corner 10 minutes into the second half. Chelsea, who held a 2-1 aggregate lead from the first leg, were second best in the first half and there were some boos at the end of an opening 45 minutes in which the home side did not have a shot.
---
Xavi Simons scored his seventh Bundesliga goal this season, having netted eight while on loan from Paris Saint-Germain last term
1: Xavi Simons scored his seventh Bundesliga goal this season, having netted eight while on loan from Paris Saint-Germain last term X
RB Leipzig moved up to fifth place in the Bundesliga with a 2-0 win over Champions League quarter finalists Borussia Dortmund, who struck the woodwork three times in a pulsating second half. Leipzig looked to be coasting after dominating the first 50 minutes and going ahead through goals by Xavi Simons and Lois Openda. But Dortmund responded strongly, with Maximilian Beier hitting the bar and the post within five minutes of Leipzig's second goal.
---
FT: Rotherham 2-3 Wycombe
2: FT: Rotherham 2-3 Wycombe
Still dazzled by the conclusion to that match? Us too. You'll have plenty of time to recover though as EFL action takes a short break - but we do mean short.
---
Jamal Musiala has signed a new contract that will keep him at Bayern Munich until at least 2030
3: Jamal Musiala has signed a new contract that will keep him at Bayern Munich until at least 2030
"People go to the stadium because of Jamal Musiala. He is an extraordinary player. Sometimes, you have the feeling gravity doesn't apply to him." That was the praise for the 21-year-old from Bayern Munich president Herbert Hainer as the club announced Musiala had extended his contract until 2030.
---
Wilson Odobert (right) scored his first two goals for Tottenham since a summer move from Burnley
4: Wilson Odobert (right) scored his first two goals for Tottenham since a summer move from Burnley X
Wilson Odobert scored twice to save Tottenham Hotspur's season as they overcame a first-leg deficit to beat AZ Alkmaar and reach the Europa League quarter-finals. Sitting 13th in the Premier League and out of both domestic cups, this competition is Spurs' only remaining chance of success. They were poor in the first leg last week in the Netherlands, losing 1-0, but much improved at home on Thursday.
---
Jayden Bogle scored Leeds' equaliser early in the second half
5: Jayden Bogle scored Leeds' equaliser early in the second half X
Championship leaders Leeds United fought back from two goals down to earn a draw at Queens Park Rangers. Koki Saito and captain Steve Cook put the hosts in command but Leeds halved their deficit shortly before half-time through a Morgan Fox own goal. Full-back Jayden Bogle equalised early in the second half as Leeds were inspired by the half-time introduction of Willy Gnonto.
---
Leroy Sane (right) scored his seventh league goal of the season to put Bayern ahead
6: Leroy Sane (right) scored his seventh league goal of the season to put Bayern ahead X
Bayern Munich failed to win for a second consecutive Bundesliga game as they were held to a draw at Union Berlin. After toiling for 75 minutes, Bayern appeared to have found the win to move them 11 points clear at the top through Leroy Sane. But Benedict Hollerbach equalised eight minutes later to earn a point for Union and move them further clear of the relegation zone.
---
Patrick Dorgu put in an impressive performance against Real Sociedad
7: Patrick Dorgu put in an impressive performance against Real Sociedad X
It was a remarkable display of sportsmanship. With Manchester United leading their Europa League last-16 tie with Real Sociedad 2-1 on the night and 3-2 on aggregate, the Red Devils were awarded a penalty. It was their third spot-kick of the match and came as referee Benoit Bastien felt Patrick Dorgu had been brought down by Hamari Traore.
---
Konstantinos Karetsas has switched allegiance from Belgium to Greece
8: Konstantinos Karetsas has switched allegiance from Belgium to Greece
Scotland take a hard-earned 1-0 lead into Sunday's second leg at Hampden Park, but no-one in their ranks is taking overall victory for granted. Greece were a much-improved side in the second half in Athens, in no small part thanks to the half-time introduction of 17-year-old Konstantinos Karetsas. The teenager from Genk in Belgium was given his debut by manager Ivan Jovanovic and immediately helped swing the momentum in the home side's favour.
---
data/human/wiki.txt.txt → data/human/wiki.txt
View file @ 9b280edb
wiki.txt Extracts:
Schrödinger's_cat
0: Schrödinger's_cat X
In quantum mechanics, Schrödinger's cat is a thought experiment concerning quantum superposition. In the thought experiment, a hypothetical cat in a closed box may be considered to be simultaneously both alive and dead while it is unobserved, as a result of its fate being linked to a random subatomic event that may or may not occur. This experiment, viewed this way, is described as a paradox.
---
Entropy
1: Entropy X
Entropy is a scientific concept, most commonly associated with states of disorder, randomness, or uncertainty. The term and the concept are used in diverse fields, from classical thermodynamics, where it was first recognized, to the microscopic description of nature in statistical physics, and to the principles of information theory. It has found far-ranging applications in chemistry and physics, in biological systems and their relation to life, in cosmology, economics, sociology, weather science, climate change and information systems including the transmission of information in telecommunication.
---
Theory_of_relativity
2: Theory_of_relativity
The theory of relativity usually encompasses two interrelated physics theories by Albert Einstein: special relativity and general relativity, proposed and published in 1905 and 1915, respectively. Special relativity applies to all physical phenomena in the absence of gravity. General relativity explains the law of gravitation and its relation to the forces of nature.
---
DNA_replication
3: DNA_replication X
In molecular biology, DNA replication is the biological process of producing two identical replicas of DNA from one original DNA molecule. DNA replication occurs in all living organisms acting as the most essential part of biological inheritance. This is essential for cell division during growth and repair of damaged tissues, while it also ensures that each of the new cells receives its own copy of the DNA.
---
Artificial_intelligence
4: Artificial_intelligence
Artificial intelligence (AI) refers to the capability of computational systems to perform tasks typically associated with human intelligence, such as learning, reasoning, problem-solving, perception, and decision-making. It is a field of research in computer science that develops and studies methods and software that enable machines to perceive their environment and use learning and intelligence to take actions that maximize their chances of achieving defined goals. Such machines may be called AIs.
---
Global_warming
5: Global_warming X
Present-day climate change includes both global warming—the ongoing increase in global average temperature—and its wider effects on Earth’s climate system. Climate change in a broader sense also includes previous long-term changes to Earth's climate. The current rise in global temperatures is driven by human activities, especially fossil fuel burning since the Industrial Revolution.
---
Evolutionary_biology
6: Evolutionary_biology X
Evolutionary biology is the subfield of biology that studies the evolutionary processes such as natural selection, common descent, and speciation that produced the diversity of life on Earth. In the 1930s, the discipline of evolutionary biology emerged through what Julian Huxley called the modern synthesis of understanding, from previously unrelated fields of biological research, such as genetics and ecology, systematics, and paleontology. The investigational range of current research has widened to encompass the genetic architecture of adaptation, molecular evolution, and the different forces that contribute to evolution, such as sexual selection, genetic drift, and biogeography.
---
Quantum_entanglement
7: Quantum_entanglement
Quantum entanglement is the phenomenon of a group of particles being generated, interacting, or sharing spatial proximity in a manner such that the quantum state of each particle of the group cannot be described independently of the state of the others, including when the particles are separated by a large distance. The topic of quantum entanglement is at the heart of the disparity between classical physics and quantum physics: entanglement is a primary feature of quantum mechanics not present in classical mechanics. Measurements of physical properties such as position, momentum, spin, and polarization performed on entangled particles can, in some cases, be found to be perfectly correlated. For example, if a pair of entangled particles is generated such that their total spin is known to be zero, and one particle is found to have clockwise spin on a first axis, then the spin of the other particle, measured on the same axis, is found to be anticlockwise.
---
Blockchain
8: Blockchain X
A blockchain is a distributed ledger with growing lists of records (blocks) that are securely linked together via cryptographic hashes. Each block contains a cryptographic hash of the previous block, a timestamp, and transaction data (generally represented as a Merkle tree, where data nodes are represented by leaves). Since each block contains information about the previous block, they effectively form a chain (compare linked list data structure), with each additional block linking to the ones before it.
---
src/__pycache__/automatic_metrics.cpython-312.pyc 0 → 100644
View file @ 9b280edb
File added
src/__pycache__/text_extraction.cpython-312.pyc
View file @ 9b280edb
No preview for this file type
src/automatic_metrics.py
View file @ 9b280edb
......@@ -16,6 +16,8 @@ def count_syllables(word):
syllable_count -= 1
return max(syllable_count, 1)
class Compute_Metrics(object):
def __init__(self):
......@@ -74,8 +76,9 @@ class Compute_Metrics(object):
pmi = math.log2(p_w1_w2 / (p_w1 * p_w2))
pmi_scores[word1, word2] = pmi
sorted_pmi_scores = dict(sorted(pmi_scores.items(), key=lambda item: item[1], reverse=True))
return sorted_pmi_scores
# sorted_pmi_scores = dict(sorted(pmi_scores.items(), key=lambda item: item[1], reverse=True))
avg_pmi = sum(pmi_scores.values()) / len(pmi_scores)
return avg_pmi
def compute_tfidf(self, new_text) -> dict[str, float]:
corpus = copy.deepcopy(self.corpus_strings)
......@@ -91,12 +94,12 @@ class Compute_Metrics(object):
feature_indices = [i for i, word in enumerate(feature_names) if word in new_text_tokens]
tfidf_dict = {feature_names[i]: tfidf_scores[i] for i in feature_indices}
sorted_tfidf = dict(sorted(tfidf_dict.items(), key=lambda item: item[1], reverse=True))
return sorted_tfidf
# sorted_tfidf = dict(sorted(tfidf_dict.items(), key=lambda item: item[1], reverse=True))
avg_tfidf = sum(tfidf_dict.values()) / len(tfidf_dict)
return avg_tfidf
def main():
test_text = """I have been dreaming of this day for years, when the sky is blue. But now I'm tired enough to see it coming up out there in my dreams as if they were some kind thing that was born from nothing but sunlight…
test_text = """I have been dreaming of this day for years, when the sky is blue. But now I\'m tired enough to see it coming up out there in my dreams as if they were some kind thing that was born from nothing but sunlight… It\'s so bright here at night because you\'re awake right next door; just like before! The sun goes down on your back while every single other part has fallen into place by then – all over me with its own little circle around us."
"""
Corpus = Compute_Metrics()
pmi = Corpus.compute_pmi(test_text)
......
src/evaluate_automatic.py 0 → 100644
View file @ 9b280edb
import automatic_metrics as am
from pathlib import Path
import re
import copy
def extract_marked_text(file_path):
"""Uses "X" as a marker to find which lines/texts to extract. Passages without a X are ignored."""
with open(file_path, 'r', encoding='utf-8') as file:
lines = file.readlines()
final_poems = {}
poems = []
current_poem = []
collecting = False
idx = 0
for line in lines:
match = re.match(r"(\d+):.*\bX\s*$", line.strip())
if collecting and current_poem:
if current_poem and any(line.strip() for line in current_poem):
poems.append("\n".join(current_poem).strip())
current_poem = []
if match:
if collecting:
final_poems[idx] = ''.join(poems)
idx += 1
poems = []
collecting = True
elif re.match(r"\d+:", line.strip()):
if collecting:
final_poems[idx] = ''.join(poems)
idx += 1
poems = []
collecting = False
elif collecting:
current_poem.append(line.strip())
if collecting and current_poem:
poems.append("\n".join(current_poem).strip())
if collecting and current_poem:
final_poems[idx] = ''.join(poems)
return final_poems
def get_all_data_from_folder(foldername, datatype="txt"):
"""extracts all files from given folder for further processing"""
script_dir = Path(__file__).resolve().parent
data_dir = script_dir.parent / f"{foldername}"
files = list(data_dir.rglob(f"*.{datatype}"))
all_extracted_text = {}
for file in files:
relativ_file_location = file.relative_to(data_dir)
text = extract_marked_text(file)
all_extracted_text[str(relativ_file_location)] = text
return all_extracted_text
def calculate_scores_texts(text):
"""Calculates scores for given text"""
texts = copy.deepcopy(text)
evaluator = am.Compute_Metrics()
evaluated_texts = {}
for filename in texts:
for idx in texts[filename]:
text = texts[filename][idx]
calc_metrics = []
calc_metrics.append(evaluator.compute_fre(text))
calc_metrics.append(evaluator.compute_ttr(text))
calc_metrics.append(evaluator.compute_pmi(text))
calc_metrics.append(evaluator.compute_tfidf(text))
evaluated_texts[f"{filename}\\{idx}"] = calc_metrics
return evaluated_texts # {filename\idx: [fre, ttr, pmi, tfidf]}
class Calculate_Parameters(object):
""""automated procedure to calculate parameters"""
def __init__(self, metrics_ai, metrics_human, question_num):
# FRE, TTR, PMI, and TF-IDF are at index 0, 1, 2, and 3
self.ai_fre, self.ai_ttr, self.ai_pmi, self.ai_tfidf = metrics_ai[0], metrics_ai[1], metrics_ai[2], metrics_ai[3]
self.human_fre, self.human_ttr, self.human_pmi, self.human_tfidf = metrics_human[0], metrics_human[1], metrics_human[2], metrics_human[3]
self.question_num = question_num
def calculate_coherence(self):
score = 0
if abs(self.ai_fre - self.human_fre) >= 20:
if self.ai_fre > self.human_fre:
score += 1
else:
score -= 1
if abs(self.ai_pmi - self.human_pmi) >= 0.8:
if self.ai_pmi > self.human_pmi:
score += 2
else:
score -= 2
if abs(self.ai_tfidf - self.human_tfidf) >= 0.2:
if self.ai_tfidf > self.human_tfidf:
score += 1
elif self.ai_tfidf < self.human_tfidf:
score -= 1
if score > 0:
return "ai"
if score < 0:
return "human"
if score == 0:
return "equal"
def calculate_conciseness(self):
score = 0
if abs(self.ai_pmi - self.human_pmi) >= 1:
if self.ai_pmi > self.human_pmi:
score += 1
else:
score -= 1
if abs(self.ai_ttr - self.human_ttr) >= 0.1:
if self.ai_ttr < self.human_ttr:
score += 2
elif self.ai_ttr > self.human_ttr:
score -= 2
if score > 0:
return "ai"
if score < 0:
return "human"
if score == 0:
return "equal"
def calculate_creativity(self):
score = 0
if abs(self.ai_pmi - self.human_pmi) >= 1:
if self.ai_pmi < self.human_pmi:
score += 1
else:
score -= 1
if abs(self.ai_ttr - self.human_ttr) >= 0.1:
if self.ai_ttr > self.human_ttr:
score += 1
elif self.ai_ttr < self.human_ttr:
score -= 1
if abs(self.ai_fre - self.human_fre) >= 20:
if self.ai_fre < self.human_fre:
score += 1
else:
score -= 1
if score > 0:
return "ai"
if score < 0:
return "human"
if score == 0:
return "equal"
def calculate_clarity_of_concept(self):
score = 0
if abs(self.ai_pmi - self.human_pmi) >= 1:
if self.ai_pmi < self.human_pmi:
score += 1
else:
score -= 1
if abs(self.ai_ttr - self.human_ttr) >= 0.1:
if self.ai_ttr < self.human_ttr:
score += 1
else:
score -= 1
if abs(self.ai_fre - self.human_fre) >= 20:
if self.ai_fre < self.human_fre:
score += 1
else:
score -= 1
if abs(self.ai_tfidf - self.human_tfidf) >= 0.2:
if self.ai_tfidf > self.human_tfidf:
score += 1
elif self.ai_tfidf < self.human_tfidf:
score -= 1
if score > 0:
return "ai"
if score < 0:
return "human"
if score == 0:
return "equal"
def predict_human_ai(survey_assessment):
"""
Counts the times when the human text had better scores on the parameters and count the times when the
ai text had better scores on the parameters.
The outputed tag is the predicted tag. "Equal" means it couldn't decide.
"""
predicted_tags = {}
for question_num, rated_param in survey_assessment.items():
ai = 0
human = 0
keys_to_check = set(rated_param.keys())
if question_num <= 6:
keys_to_check.discard("clarity_of_concept")
elif 6 < question_num <= 12:
keys_to_check.discard("creativity")
elif 12 < question_num <= 18:
keys_to_check.discard("clarity_of_concept")
keys_to_check.discard("creativity")
for key in keys_to_check:
if rated_param[key] == "ai":
ai += 1
elif rated_param[key] == "human":
human += 1
# Determine result
if human > ai:
predicted_tags[question_num] = "human"
elif human < ai:
predicted_tags[question_num] = "ai"
else:
predicted_tags[question_num] = "equal"
return predicted_tags
if __name__ == '__main__':
survey_texts = get_all_data_from_folder("data", "txt")
evaluated_texts = calculate_scores_texts(survey_texts)
# I manually ordered the texts in the order used in the survey
survey_ai_texts = ['ai\\gpt2_poem.txt\\0', 'ai\\gpt2_poem.txt\\1', 'ai\\opt_poem.txt\\0', 'ai\\opt_poem.txt\\1', 'ai\\gpt4o_poem.txt\\0', 'ai\\gpt4o_poem.txt\\1',
'ai\\gpt4o_wiki.txt\\0', 'ai\\gpt4o_wiki.txt\\1', 'ai\\opt_wiki.txt\\0', 'ai\\opt_wiki.txt\\1', 'ai\\gpt2_wiki.txt\\0', 'ai\\gpt2_wiki.txt\\1',
'ai\\opt_sport.txt\\0', 'ai\\opt_sport.txt\\1', 'ai\\gpt4o_sports.txt\\0', 'ai\\gpt4o_sports.txt\\1', 'ai\\gpt2_sport.txt\\0', 'ai\\gpt2_sport.txt\\1'
]
survey_human_texts = ["human\\poetry.txt\\0", 'human\\poetry.txt\\1', 'human\\poetry.txt\\2', 'human\\poetry.txt\\3', 'human\\poetry.txt\\4', 'human\\poetry.txt\\5',
'human\\wiki.txt\\0', 'human\\wiki.txt\\1', 'human\\wiki.txt\\2', 'human\\wiki.txt\\3', 'human\\wiki.txt\\4', 'human\\wiki.txt\\5',
'human\\sport_bbc.txt\\0', 'human\\sport_bbc.txt\\1', 'human\\sport_bbc.txt\\2', 'human\\sport_bbc.txt\\3', 'human\\sport_bbc.txt\\4', 'human\\sport_bbc.txt\\5'
]
survey_groups = zip(survey_ai_texts, survey_human_texts)
# Rate parameters Coherence, Creativity, Conciseness, Clarity of Concepts between survey groups
survey_assessment = {}
for i, group in enumerate(survey_groups, start=1):
for idx, name in enumerate(group):
if idx < len(group) - 1:
metrics_ai = evaluated_texts[name]
metrics_human = evaluated_texts[group[idx + 1]]
evaluation_metrics = Calculate_Parameters(metrics_ai, metrics_human, i)
coherence_score = evaluation_metrics.calculate_coherence()
conciseness_score = evaluation_metrics.calculate_conciseness()
creativity_score = evaluation_metrics.calculate_creativity()
clarity_score = evaluation_metrics.calculate_clarity_of_concept()
survey_assessment[i] = {
"coherence": coherence_score,
"conciseness": conciseness_score,
"creativity": creativity_score,
"clarity_of_concept": clarity_score
}
# Automatically asses if text is human or ai genareted
result = predict_human_ai(survey_assessment)
src/text_extraction.py
View file @ 9b280edb
......@@ -114,7 +114,7 @@ if __name__ == "__main__":
# output_directory = r"C:\Users"
# write_file(output_directory, "poetry_newlines.txt", poetry_text[0])
# write_file(output_directory, "wiki.txt", wiki_text)
# write_file(output_directory, "sport_bbc.txt", sports_text)
# write_file(output_directory, "poetry_newlines", poetry_text[0])
# write_file(output_directory, "wiki", wiki_text)
# write_file(output_directory, "sport_bbc", sports_text)