Datasets:

cirimus
/

super-emotion

text stringlengths 1 830	labels listlengths 1 5	labels_str listlengths 1 5	labels_source listlengths 1 6	source stringclasses 10 values
also I was the point person on my companys transition from the KL-5 to GR-6 system.	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
You mustve had your hands full.	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
That I did. That I did.	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
So lets talk a little bit about your duties.	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
My duties? All right.	[ 6 ]	[ "Surprise" ]	[ "surprise" ]	MELD
Now youll be heading a whole division, so youll have a lot of duties.	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
I see.	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
But therell be perhaps 30 people under you so you can dump a certain amount on them.	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
Good to know.	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
We can go into detail	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
No dont I beg of you!	[ 1 ]	[ "Fear" ]	[ "fear" ]	MELD
All right then, well have a definite answer for you on Monday, but I think I can say with some confidence, youll fit in well here.	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
Absolutely. You can relax	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
But then who? The waitress I went out with last month?	[ 6 ]	[ "Surprise" ]	[ "surprise" ]	MELD
You know? Forget it!	[ 5 ]	[ "Sadness" ]	[ "sadness" ]	MELD
No-no-no-no, no! Who, who were you talking about?	[ 6 ]	[ "Surprise" ]	[ "surprise" ]	MELD
No, I-I-I-I don't, I actually don't know	[ 1 ]	[ "Fear" ]	[ "fear" ]	MELD
Ok!	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
All right, well...	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
Yeah, sure!	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
Hey, Mon.	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
Hey-hey-hey. You wanna hear something that sucks.	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
Do I ever.	[ 2 ]	[ "Joy" ]	[ "joy" ]	MELD
Chris says theyre closing down the bar.	[ 5 ]	[ "Sadness" ]	[ "sadness" ]	MELD
Yeah, apparently theyre turning it into some kinda coffee place.	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
Just coffee! Where are we gonna hang out now?	[ 0 ]	[ "Anger" ]	[ "disgust" ]	MELD
Can I get a beer.	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
Hey, did you pick a roommate?	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
You betcha!	[ 2 ]	[ "Joy" ]	[ "joy" ]	MELD
Is it the Italian guy?	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
Um-mm, yeah right!	[ 2 ]	[ "Joy" ]	[ "joy" ]	MELD
Oh my God, oh my God! Poor Monica!	[ 6 ]	[ "Surprise" ]	[ "surprise" ]	MELD
What, what, what?!	[ 6 ]	[ "Surprise" ]	[ "surprise" ]	MELD
He was with her when he wrote this poem.	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
Look, 'My vessel so empty with nothing inside.	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
Now that I've touched you, you seem emptier still.'	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
He thinks Monica is empty, she is the empty vase!	[ 6 ]	[ "Surprise" ]	[ "surprise" ]	MELD
Oh, totally. Oh, God, oh, she seemed so happy too.	[ 5 ]	[ "Sadness" ]	[ "sadness" ]	MELD
Done.	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
Ah y'know, this building is on my paper route so I...	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
Howd did it go?	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
Oh well, the woman I interviewed with was pretty tough, but y'know thank God Mark coached me, because once I started talking about the fall line, she got all happy and wouldnt shut up.	[ 2 ]	[ "Joy" ]	[ "joy" ]	MELD
Im so proud of you.	[ 2 ]	[ "Joy" ]	[ "joy" ]	MELD
Listen, Im ah, Im sorry Ive been so crazy and jealous and, its just that I like you a lot, so...	[ 5 ]	[ "Sadness" ]	[ "sadness" ]	MELD
Ameri-can.	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
Ameri-ccan.	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
Ameri-can. Y'know its a	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
Everybody!!	[ 6 ]	[ "Surprise" ]	[ "surprise" ]	MELD
Good job Joe! Well done! Top notch!	[ 2 ]	[ "Joy" ]	[ "joy" ]	MELD
You liked it? You really liked it?	[ 6 ]	[ "Surprise" ]	[ "surprise" ]	MELD
Oh-ho-ho, yeah!	[ 2 ]	[ "Joy" ]	[ "joy" ]	MELD
Which part exactly?	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
The whole thing! Can we go?	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
Oh no-no-no, give me some specifics.	[ 0 ]	[ "Anger" ]	[ "anger" ]	MELD
I love the specifics, the specifics were the best part!	[ 2 ]	[ "Joy" ]	[ "joy" ]	MELD
Hey, what about the scene with the kangaroo? Did-did you like that part?	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
I was surprised to see a kangaroo in a World War I epic.	[ 6 ]	[ "Surprise" ]	[ "surprise" ]	MELD
You fell asleep!!	[ 0 ]	[ "Anger" ]	[ "anger" ]	MELD
There was no kangaroo!	[ 0 ]	[ "Anger" ]	[ "anger" ]	MELD
They didnt take any of my suggestions!	[ 0 ]	[ "Anger" ]	[ "anger" ]	MELD
Thats for coming buddy.	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
Ill see you later.	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
Dont go!	[ 5 ]	[ "Sadness" ]	[ "sadness" ]	MELD
Im so sorry!	[ 5 ]	[ "Sadness" ]	[ "sadness" ]	MELD
This guy fell asleep!	[ 0 ]	[ "Anger" ]	[ "anger" ]	MELD
He fell asleep too!	[ 0 ]	[ "Anger" ]	[ "anger" ]	MELD
Be mad at him!	[ 0 ]	[ "Anger" ]	[ "anger" ]	MELD
Or, call an ambulance.	[ 0 ]	[ "Anger" ]	[ "anger" ]	MELD
Okay, look, I think we have to tell Rachel she messed up her dessert.	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
What?! What is with everybody? Its Thanksgiving, not...Truth-Day!	[ 0 ]	[ "Anger" ]	[ "anger" ]	MELD
Yes, and it is my dying wish to have that ring.	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
See, if Im not buried with that ring then my spirit is going to wander the nether world for all eternity	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
Okay, thats enough honey!	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
I dont know. Let me see the ring.	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
Great! Okay, here.	[ 2 ]	[ "Joy" ]	[ "joy" ]	MELD
Thank you. Thank you. Thank you! And	[ 2 ]	[ "Joy" ]	[ "joy" ]	MELD
What've you been up to?	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
Oh, you know, the usual, teaching aerobics, partying way too much.	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
Oh, and in case you were wondering, those are my legs on the new James Bond poster.	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
Can you hold on a moment? I have another call. I love her.	[ 2 ]	[ "Joy" ]	[ "joy" ]	MELD
I'm back.	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
So, are we gonna get together or what?	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
Um, absolutely. Uh, how 'bout tomorrow afternoon? Do you know uh, Central Perk in the Village, say, five-ish?	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
Great, I'll see you then.	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
Ok. Ok. Having a phone has finally paid off.	[ 2 ]	[ "Joy" ]	[ "joy" ]	MELD
Even though you do do a good Bob impression, I'm thinkin' when she sees you tomorow, she's probably gonna realize, "Hey, you're not Bob."	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
I'm hoping that when Bob doesn't show up, she will seek comfort in the open arms of the wry stranger at the next table.	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
Oh my god. You are pure evil.	[ 0 ]	[ "Anger" ]	[ "disgust" ]	MELD
Ok, pure evil , horny and alone . I've done this	[ 0 ]	[ "Anger" ]	[ "anger" ]	MELD
Nobody!	[ 0 ]	[ "Anger" ]	[ "anger" ]	MELD
Nobody respects the bucket!	[ 0 ]	[ "Anger" ]	[ "anger" ]	MELD
You wouldn't believe what people put in here!	[ 0 ]	[ "Anger" ]	[ "anger" ]	MELD
Okay, does this look like a garbage can to you?	[ 0 ]	[ "Anger" ]	[ "anger" ]	MELD
This is my wedding.	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
All right, yknow what? We really are late! Lets go! Lets go! Lets go!!	[ 0 ]	[ "Anger" ]	[ "anger" ]	MELD
Fine. Youll-youll watch it on video when we get back.	[ 4 ]	[ "Neutral" ]	[ "neutral" ]	MELD
Here you go Phoebe! Here you go Phobo! Phewbedo! Phaybobo.	[ 2 ]	[ "Joy" ]	[ "joy" ]	MELD
All right, lets go! Bye, Pheebs!	[ 2 ]	[ "Joy" ]	[ "joy" ]	MELD
Bye, Pheebs!	[ 2 ]	[ "Joy" ]	[ "joy" ]	MELD
London baby!!	[ 2 ]	[ "Joy" ]	[ "joy" ]	MELD

End of preview. Expand in Data Studio

Super Emotion Dataset

Associated Paper

We provide full documentation of the dataset construction process in the accompanying paper: 📘 The Super Emotion Dataset (PDF)

Dataset Summary

The Super Emotion dataset is a large-scale, multilabel dataset for emotion classification, aggregated from five prominent emotion datasets:

It contains 507,805 unique text samples and 535,227 total emotion label assignments, reflecting its multilabel nature. Emotion labels are mapped into six primary emotions and a neutral class: joy, sadness, anger, fear, love, surprise, and neutral. Related or ambiguous categories like happiness, hate, or grief are merged accordingly.

Supported Tasks

This dataset is designed for emotion classification and can be used for:

Multi-label emotion recognition
Emotion co-occurrence modeling
Single-label classification
Fine-tuning language models for affective NLP

Dataset Structure

The dataset follows the structure:

Column	Type	Description
text	string	The raw input text
labels	list of int	One or more mapped emotion labels (ids in `label2id.json`)
labels_str	list of string	The same labels as names
labels_source	list of string	The original labels in the source dataset
source	string	The original dataset name

Splits:

Train: 407,186 samples
Validation: 48,669 samples
Test: 51,950 samples

Class distribution by source:

Source	Neutr.	Surp.	Fear	Sad.	Joy	Anger	Love	Total
MELD	5662	1311	344	951	2042	1854	0	12164
CARER	0	14075	46284	120138	138846	56447	32809	408599
GoEmotions	17711	4281	926	4019	7609	7818	15417	57781
Crowdflower	8781	2182	8442	5154	9237	1432	3792	39020
SemEval	0	566	1848	3607	5065	4779	1757	17622
Several	15	7	0	4	6	3	6	41
All	32169	22422	57844	133873	162805	72333	53781	535227

License

Super Emotion is a compilation: every text keeps the licence and terms of its source dataset. CARER, about 78% of the rows, may be used for educational and research purposes only, so the dataset as a whole is for non-commercial research and education.

Source	Licence or terms
CARER	educational and research use only
GoEmotions	Apache 2.0
MELD	GPL-3.0
SemEval	the task's terms and conditions
Crowdflower	no licence stated by the source

The source column says where each text comes from, so a subset can be selected by source. Check the licence of each source before redistribution or commercial use.

Citation

If you use this dataset, please cite both the original sources and this aggregation:

@article{junquedefortuny2025superemotion,
  title   = {The Super Emotion Dataset},
  author  = {Enric Junqu{\'e} de Fortuny},
  journal = {arXiv preprint},
  year    = {2025},
  url     = {https://arxiv.org/abs/2505.15348},
  note    = {Dataset and supplementary materials available at \url{https://huggingface.co/datasets/cirimus/super-emotion}}
}

Versions

Version	Date	Changes
v2	2026-10	"ISEAR" relabelled (CARER), some train/test leakage removed
v1	2025-05	First release

Downloads last month: 312

Models trained or fine-tuned on cirimus/super-emotion

Spaces using cirimus/super-emotion 2

Paper for cirimus/super-emotion

The Super Emotion Dataset

Paper • 2505.15348 • Published May 21, 2025